KR100802511B1 - System and method for offering searching service based on topics - Google Patents

Abstract

The present invention relates to an Internet search service system and a method thereof. To this end, the present invention, unlike the prior art that does not provide an identification method for a homonym and a synonym of a keyword included in a search query, uses a TM (Topic Manager) Registration request signal, TP (Topic Page) registration request signal, content for a specific topic, registration request signal for a new topic, new synonym addition request signal for a specific topic, hyperlink link request signal for another TP, search query A client that transmits one of a request signal and a user additional function selection signal to a search service server through a wired / wireless network, and performs a signal processing for each request signal, and topics related to the search query according to the search request signal for the search query. List, TP summary list registered to the selected topic on the client, TP either By constructing a search service providing system that includes a search service server that transmits to a client through a trust network, the meaning of the user's search query by constructing a topic-based search service providing system that can identify search queries, homonyms, and synonyms It can effectively present the search results that match the.

Search query, topic, search, TM, TP, hyperlink

Description

Topic-based search service providing system and method thereof {SYSTEM AND METHOD FOR OFFERING SEARCHING SERVICE BASED ON TOPICS}

1 is a block diagram of a topic-based search service providing system suitable for providing a search service through a TP created based on a topic according to the present invention;

2 is a flowchart illustrating a process of registering a TM according to a TM registration request signal according to the present invention;

3 is a flowchart illustrating a process of registering a TP created after user authentication and registering a hyperlink connection according to the present invention;

4 is a diagram illustrating a search homepage according to the present invention;

5 illustrates a TP editing tool web page in accordance with the present invention;

6 illustrates an email regarding a TP registration review and reply request in accordance with the present invention;

7 is a diagram illustrating an email in the form of a survey request of a TP registration review request according to the present invention;

8 is a diagram illustrating a TP registration review request response status table according to the present invention;

9 is a diagram illustrating a relationship between an index word, a topic, a synonym list, and a TP according to the present invention;

10 is a diagram illustrating a relationship according to another embodiment between an index word, a topic, a synonym list, and a TP according to the present invention;

11 is a diagram illustrating a detailed structure of a topic related information list according to the present invention;

12 is a diagram illustrating a hyperlink connection request information input window according to the present invention;

13 is a flowchart illustrating a process of searching for a TP created based on a topic according to the present invention;

14 is a diagram illustrating a function of setting a TP search condition according to the present invention;

FIG. 15 illustrates a first search results webpage displaying topics, TP summaries, TP detection, and TP display via a TP search in accordance with the present invention; FIG.

16 illustrates a third search result web page for detecting TPs and displaying TPs via a TP search in accordance with the present invention;

17 is a flowchart illustrating a process of performing a process of processing a hyperlink selection signal included in a TP search result according to the present invention;

18 is a diagram illustrating a hyperlink related information table between TPs according to the present invention;

19 is a flowchart illustrating a process of performing a process of an optional user add-on selected through a TP search result according to the present invention;

20 is a diagram illustrating various forms of a user evaluation score display and an input interface for a user evaluation function for a TP among user additional functions according to the present invention;

21 is a diagram illustrating a storage structure of user rating popularity-related data according to the present invention;

FIG. 22 is a flowchart illustrating a process of processing a batch reproduction TP ranking value element including user evaluation reliability at an evaluation revalidation time point according to the present invention; FIG.

23 is a diagram illustrating a storage structure of a user evaluation score grant situation for each TP according to the present invention;

24 is a flowchart illustrating a process of additionally registering a new topic according to the present invention;

25 is a flowchart illustrating a process of additionally registering a topic synonym according to the present invention;

26 is a view illustrating a review request email when a topic synonym addition registration request according to the present invention;

27 illustrates a review request email when requesting to add a synonym for a designated topic in accordance with the present invention.

28A and 28B are flowcharts illustrating a process of providing a search service based on a topic according to the present invention;

<Description of the symbols for the main parts of the drawings>

102: client group 102a: TM client

102b: Competitive TM Client 102c: HyperlinkTM Client

102d: Mentor Client 102e: Search User Client

102f: Operator Client 104: Wired or Wireless Network

106: search service server 106a: TP processing unit

106b: TP search unit 106c: topic processing unit

106d: TM processing unit 106e: Hyperlink processing unit

106f: evaluation processing unit 106g: authentication processing unit

106h: control unit 108: storage device

108a: Index structure database for index word-topic mapping

108b: Index structure database for topic-TP mapping

108c: Index structure database for indexer-TP mapping

108d: Topic Information Database 108e: TP Information Database

108f: TM Information Database 108g: Hyperlink Database

The present invention relates to a technique for providing an Internet search service, and more particularly, to a topic-based search service providing system and method suitable for providing a topic-based Internet search service, which is an identification unit of a homonym and a synonym. will be.

As is well known, the search engines of Internet search portals such as Yahoo (www.yahoo.com), Naver (www.naver.com), Empas (www.empas.com), Google (www.google.com), etc. It visits a server where a web server on a network called a robot (aka Crawler) is installed, uses a software to collect web documents, makes a database of collected web documents, and helps users search.

Such search engines automatically index encyclopedias, images and news, which are difficult to collect on the web, and provide search services by integrating with web document search services to search users. Here, the search user requests a search by inputting a keyword or a search query.

Then, the search information for the search query entered is firstly summarized information (eg, url, title summary, thumbnail, date, ranking score, etc.) ranked according to the match with the search query entered by the user. It shows a list of, and secondly you can browse the detailed information content of the selected item in the list.

However, search engines of Internet search portals such as Yahoo, Naver, Empas, and Google have the following problems. First, the problem of discernment of homonyms (for example, CD has two meanings: compact disk and an anonymous deposit certificate), and second, the problem of recognition of the synonym (for example, CD and compact discs have different forms). Third, the problem of notation of various Korean pronunciations of foreign words (e.g., similar problem with the recognition of a synonym with different Korean notation such as digital, digital, digital, etc.) Problem of uniformity and inadequacy of ranking evaluation (for example, lack of understanding of semantic theme of documents by ranking decision based on keyword's location in documents, frequency-based statistical weighting, hyperlink weighting, etc.) Incompleteness issues (e.g. missing collection of documents or information of various types and sources due to limitations in web document collection cycles, site access restrictions, etc.) There are problems such as lack of abundance and accuracy of the color main eel closely related recommendations.

As an alternative to these problems, we have attempted to solve the above problems by building synonym dictionaries, thesauruses, etc. in large quantities, but there are many errors in the construction data itself due to limited manpower, resources, period, etc. Lack of information on the terminology of the inevitability is inevitably low.

In addition, in the search portal (www.daum.net), we are trying to access these problems using 'SearchJaki' for Daum (www.daum.net) and 'Automatically suggest related words' for Empas (for example, We use statistical analysis of air relations in logs or keywords documents, built-in dictionary of related terms, etc.). However, the list of related terms presented does not satisfy the user's satisfaction because it shows relatively low search results. I can't.

The reason is that the automatic suggestion of related words uses a relatively simple statistical technique, so that less relevant data is searched. Therefore, it is not easy to improve the quality of the search results, and in the case of thesaurus, thesaurus, etc. Work on large amounts of data, including many existing data error corrections and corrections, which increases the likelihood of errors and puts a strain on ongoing tuning and the reflection of new phenomena and changes. .

In addition, large search portals attempt to approach the above-mentioned problems by artificially constructing and registering a theme-related web page that is highly related to a subject using specific dedicated workers, or by artificially adjusting the ranking of search results. For example, manual editing of search engine sites such as MSN, manual maintenance such as Yahoo! Directory, etc. These methods are performed using a limited manpower, rather than converging the views of various end user groups. It is a difficult reality to work on a large number of search terms.

In addition, a method of reducing costs and performing a large amount of work by relying on the work of expert volunteers, such as Open Directory, has been attempted, but only the categorization category and site list information of the directory are provided. The disadvantage is that there is little diversity. This Open Directory is not created by individual experts creating their own directory pages and by evaluating the views among the experts, but by jointly gathering the opinions of experts to create a single directory page.

On the other hand, there is a knowledge search service of Naver's Knowledge iN or other search portals that are being serviced as an approach to the above-mentioned problems, but provided only in the form of answers to questions, and a variety of information about the questions not raised. It is a difficult reality to accumulate search results. In addition, information related to a specific topic may be divided and provided in various places in community bulletin boards, blogs, discussion forums, etc. There was a problem that it was insufficient to provide the basis and function for searching the information related to the subject in depth contents by escaping the surface relevance including the query.

For example, services such as wikis co-edit information on specific topics and provide hyperlinking between topics, but approaches to homonyms and synonyms in search query processing are suggested. The blog service indexes keywords extracted from blog titles or texts, and searches indexes that directly match keywords (search queries). It's only a collection of sites on the site, lacks the ability to directly compare and integrate content from multiple people on the same topic, and does not provide an approach to identifying homonyms or synonyms in search. It is a reality.

Accordingly, the present invention is to solve the problems of the prior art, to provide a topic-based search service providing system and method that can effectively solve the problem of identifying homonyms or synonyms by providing search results based on the topic Its purpose is to.

Another object of the present invention is to create a topic-based TP (Topic Page) and register to one or more specific topics, to retrieve, present, and hyperlink links to other TPs that have topical relevance to TPs registered on a particular topic. The present invention provides a topic-based search service providing system and a method thereof.

It is still another object of the present invention to calculate user rating reliability based on user rating popularity for any TP and to combine various TP popularity and other ranking factors including the user rating popularity and hyperlink popularity between TPs in each search. By applying the TP ranking value of each TP that is used to determine the relative priority of the batch in the search result, a topic-based search service providing system that can improve the search quality as well as the quality and evaluation reliability of any TP. And to provide a method.

According to an aspect of the present invention, there is provided a system including a storage device and a client and a search service server connected to a wired / wireless network to provide a search service for an arbitrary search query. Request registration of TM (Topic Manager) client requesting registration of created TP (Topic Page), search user client requesting search of the search query word, and other TPs created for the same topic as the specific topic of the created TP Providing a request for the search service and a response, including a competing TM client, and processing a registration request for a TP registration request from the client, the client requesting registration of a new synonym for a new topic or a specific topic; TP processing unit, topic, TP, TP according to the search request from the client A TP retrieval unit for searching and providing any one of the drug lists, a control unit for controlling each process, and storing corresponding information in the storage device, and a new synonym for a new topic or a specific topic from the client. A topic processing unit that processes a registration request according to a registration request for the client, a TM processing unit that processes a registration request according to a TM registration request from the client, and a relative priority on a placement position of a TM summary item from the client. A topic-based search service providing system including a search service server for transmitting a response result related to a search service request from the client to the client through the wired / wireless communication network, including an evaluation processing unit calculating a TP ranking value of each TP. To provide.

According to another aspect of the present invention, there is provided a storage device, and a method in which a client and a search service server are connected through a wired or wireless network to provide a search service for an arbitrary search query. Checking whether a search request signal for a search query is received; and when the search request signal is received, a unique topic or a list of topics corresponding to the search query, and a specific topic or a registered TP corresponding to the unique topic. Generating a first search result including a summary list and transmitting the first search result to the search user client through the wired / wireless communication network; and when the topic list has two or more topics, a selection signal of a specific topic through the first search result Checking whether a signal is received and a line for the specific topic Generating a second search result including a list of TP summaries registered according to the topic when the tag signal is received, and transmitting the second search result to the search user client through the wired / wireless communication network; and the first search result or the second search result. Checking whether a summary selection signal of a specific TP is received based on the search result; when a summary selection signal of the specific TP is received, a TP corresponding to the summary of the TP; and a list of information related to hyperlinks using the TP as a source TP Generating a third search result including at least one of a selection menu related to a user additional function including a user evaluation and a user recommendation, and transmitting the third search result to the search user client through the wired / wireless communication network; During the new topic registration request signal or consent to a specific topic from the client Checking whether an additional request signal is received, and if a new topic registration request signal or a synonym addition request signal for a specific topic is received, information about the new topic registration request synonym is not registered in the storage device unless it overlaps with a previously registered topic or synonym. A topic-based search comprising a process of storing and calculating a measured value or input value of a batch recalculation TP ranking value element variable for each TP and a user evaluation reliability when a predetermined cycle occurs during the processes. Provide a service provision method.

The above and other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art.

Prior to detailed description of the present invention, terms used in the present invention will be described.

First, a topic refers to a subject to be edited, created, and registered, and is a method for coping with identification problems such as homonyms and synonyms of search queries in retrieving the contents of the information. Refers to a group of search terms including a search query or any synonym for a search query in the form of a string, phrase, clause, or sentence. The logical structure of these topics is a topic identifier, a descriptive description of the topic, and a list of synonyms ( Search term group including the synonym), and further include a higher category topic identification name. In addition, it can be connected by hyperlinks between TPs (Topic Pages) through the relationship between vocabulary, topic, and the like, and have direct or indirect connection with other topics through these connections.

The TP (Topic Page) may include a variety of information related to at least one specific topic (eg, term definitions, expertise, related sites and various web search result edits, comments, announcements, etc.), content, and programs that can be executed. As an embedded web page, it can be written in the form of a blog post, comment, remote trackback, post, or the like. And, it may be registered for one or more designated target topics. The target topic may be specified in the client, unilaterally specified in the search service server described later, or the target topic candidate may be recommended to the client.

Finally, when describing the TP ranking value element variable, the ranking value of each TP determines the relative superiority of each TP on the position placed in the list of search results for the query user's query word. Is calculated using at least one of factors such as the relationship between the TP and the query word, the TP-related metadata, the popularity of the TP itself, and the popularity or reliability of the TP author. In addition, the TP ranking value element means registration or modification time, TP file size, and number of attachments / documents in the case of TP-related metadata, and user evaluation popularity score, hyperlink popularity score, etc. it means. These TP ranking value elements are called ranking value element variables because they may have variable values derived based on measured values that are data accumulated through measurement. Here, the TP ranking value element variables are a real-time calculated TP ranking value element variable whose input values are calculated in real time at the time of the TP search according to the time point at which the input value is calculated and the calculation method, and the collective retrieval TP ranking at predetermined periodic time points. Classified as a value element variable.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The core technical aspect of the present invention, unlike the prior art that does not provide the identification method for the homonym and the synonym of the keyword included in the search query in the Internet search service system, TM registration request signal, the content for a specific topic TP (Topic Page) Search the wired / wireless network for any one of a registration request signal, a registration request signal for a new topic, a new synonym addition request signal for a specific topic, a hyperlink connection request signal for another TP, and a search request signal for a search query. A client transmitting to the service server, a signal processing for the corresponding request signal, and a list of topics related to the search query according to the search request signal for the search query, a list of TP summaries registered to the only topic corresponding to the search query, and the client. TP summary list registered on selected topic, either TP or wired / wireless network By implementing a topic-based search service providing system including a search service server transmitted to a client through the above, it is possible to easily achieve the object of the present invention through such technical means.

1 is a block diagram of a topic-based search service providing system suitable for providing a search service through a TP created based on a topic according to the present invention. In general, a client group 102 and a wired / wireless communication network 104 are provided. , Search service server 106, and storage 108. Here, the client group 102 includes a TM client 102a, a competing TM client 102b, a hyperlink TM client 102c, a mentor client 102d, a search user client 102e and an operator client 102f. The search service server 106 includes a TP processing unit 106a, a TP search unit 106b, a topic processing unit 106c, a TM processing unit 106d, a hyperlink processing unit 106e, an evaluation processing unit 106f, and an authentication processing unit ( 106g) and a controller 106h, the storage device 108 includes: an index structure database 108a for index word-topic mapping, an index structure database 108b for topic-TP mapping, an index structure database for index word-TP mapping ( 108c), topic information database 108d, TP information database 108e, TM information database 108f, and hyperlink database 108g.

Referring to FIG. 1, the TM client 102a of the client group 102 creates a TP by specifying any topic among a list of topics registered in the TP search service server, and requests index registration or storage for the specified topic. , Terminal (eg, desktop computer, PDA (Personal Digital Assistant), IMT2000, portable terminal that can be used by a user (TM: Topic Manager) who performs the task of requesting registration or storage again by modifying his registered TP) , A mobile phone, a user input / output terminal, etc.).

In addition, the competing TM client 102b registers the TPs created by the competing TM clients 102b with each other to become competing TM clients 102b to each other, and not only the competing TM client ( 102b) Terminals (eg, desktop computers, PDAs, IMT2000s, web-browsable portable devices) used by users who monitor activities such as listing abuse or violation of terms and conditions and informing TP search service operators. Terminal, mobile phone, user input / output terminal, etc.).

In addition, when the hyperlink TM client 102c establishes and establishes a hyperlink from an arbitrary TP and TP1 to another arbitrary TP and TP2, each of the TP1 and TP2 is registered and registered through the TM client 102a. Means TM clients 102a of TM2 (Hyperlink TM), wherein the two TM clients 102a become hyperlink TM clients 102c with each other.

Meanwhile, the Mentor client 102d receives the request for reviewing and approving the contents of the TP created for any topic, sends a response message thereto, and manages the activity of the TM client 102a of the TM in charge to malicious , Terminals (eg, desktop computers, PDAs, IMT2000s, web-enabled mobile terminals, mobile phones, user input / output terminals) used by users (mentors) who control the authoring and posting of unhealthy information and encourage active TP registration and editing And the like). In exchange for supervision of the responsible TM to encourage mentor activities, the search service server 106 displays the service contribution display function of the mentor's point (or mileage, air, etc.) stored in the TM information database 108f. Field), i.e. point revenue distribution, such as by adding and storing a certain percentage of points given as a reward for the contribution of the search TM to the search service over a period of time or additionally given due to an increase in popularity. The method of granting compensation of the method can be applied.

The search user client 102e indicates a terminal (eg, a portable terminal, a mobile phone, a user input / output terminal, etc.) used by a user (search user) who requests a search for a TP about a specific topic. Denotes a terminal (for example, a desktop computer, a PDA, an IMT2000, a portable terminal capable of web browsing, a mobile phone, a user input / output terminal, etc.) used by a user (operator) operating a topic-based search service providing system.

Here, the TM client 102a, the competing TM client 102b, the hyperlink TM client 102c, the mentor client 102d, the search user client 102e and the operator client 102f may be a search service server 106 or Of course, one of these clients can be recognized depending on the role played by other clients. For example, when search service server 108 receives a TP search request signal from any TM client 102a, any TM client 102a is recognized as one search user client 102e to provide a search service. do.

Next, the wired / wireless communication network 104 communicates with each client and the search service server 106, which may be a search service server and a desktop computer, a PDA, an IMT2000, a portable terminal capable of web browsing, a mobile phone, a user input / output terminal, and the like. This means a general communication network for enabling connection, for example, a wired Internet network such as TCP / IP protocol, a wireless Internet network such as WAP protocol, a mobile communication network, a LAN, and a communication line connection with a user input / output terminal device. Such a communication network may include various network / switching equipment, a web server, a load balancer, a firewall, and the like.

On the other hand, the TP processing unit 106a in the search service server 106 processes the TP registration request signal from each of the TM clients 102a or the operator client 102f, and indexes the storage in the storage device 108 for the designated topic. Index into the topic index structure database 108a and the index structure database 108b for topic-TP mapping, or the index structure database 108c for index-TP mapping, and register and store the TP requested to be registered in the TP information database 108e. And so on. Preferably, another TM client 102a that has registered a TP on a topic to be designated for registration through an SMS server, an MMS server, a mobile network, an e-mail server, or the like, requesting review and approval of the content of the TP before storing the index and registration of the TP. ) To request one to review and approve the TP registration request, and to store and index the TP when the approval is obtained.

In addition, the TP processing unit 106a registers and stores the registered points in the TP information database 108e by assigning points (or similar functions such as mileage, grade, air service, etc.) reflecting the accumulated popularity and contribution to each registered TP. And perform a coordination management function. Preferably, the topic includes a storage management function of the same TP and performs a function of transmitting a TP creation tool to each client to support TP creation of each TM clients.

Next, the TP retrieval unit 106b detects the topics corresponding to (matched) the search query (search term) received from the search user client 102e, and identifies a unique topic or a list of topics corresponding to the search query term, and the specific thereof. After generating a first search result including a topic or a list of registered TP summaries corresponding to the unique topic, the first search result is transmitted via the wired / wireless network 104 to the search user client 102e, and the topic list has two topics. In the above-described case, when a selection signal for a specific topic is received from the search user client 102e through the first search result, a summary information list of TPs registered in the selected specific topic is detected, and the registration information corresponding to the selected specific topic is registered. After generating a second search result including the TP summary list, it is sent to search user client 102e.

In addition, the TP retrieval unit 106b receives the TP corresponding to the specific TP summary information when the selection signal of the specific TP summary item is received from the search user client 102e through the first search result or the second search result. After generating a third search result including at least one of a hyperlink related information list serving as the source TP and a selection menu related to the user additional function, the third search result is transmitted to the search user client 102e.

Preferably, the TP retrieval unit 106b includes a first list including a summary list of TPs indexed to the detected index word by detecting an index word matching the search query instead of detecting a topic matching the search query (search word). You can create and send search results. The index word may be any one of a specific string, word, phrase, and sentence, and the matching method of the search query and the index word may be a complete match or a partial match matching method. In this case, perfect match means that the search query and the index match exactly, and partial match means extraction using the degree of similarity between the search query and the index. For example, a search in the search query using morphological analysis, Function words such as endings or stop words can be separated and the similarity between extracted keywords and index words can be used.

In addition, the search user client 102e may select an arbitrary user add-on through the third search result. The user add-on includes a recommendation function in which one search user recommends a TP to another search user, and a search user TP. User rating function, comment / reply function for TP content, scraping part of TP or TP content, mailing TP content to other search users, outputting TP content, trackback on blog ) And at least one of an RSS connection function. In addition, the user additional function may include a function of rearranging and displaying a list of TP summaries included in the first search result or the second search result according to other criteria such as popularity, registration date, and the like.

Next, the topic processing unit 106c receives the registration request signal of the new topic information or the synonym addition registration request signal related to the previously registered topic from each TM client or the operator client 102f to convert the new topic information into the topic information. A new synonym is registered and stored in the database 108d, or a new synonym related to a previously registered topic is added to a synonym list storage space allocated to a specific topic in the previously stored topic information database 108d. Preferably, the content of at least one of the competing TM client 102b, the mentor client 102d, and the operator client 102f of the TM client 102a requesting a new additional registration before storing the new registration or synonym's further registration. Request for review and approval, and the process of obtaining approval.

The TM processing unit 106d transmits a TM subscription information input window for the new TM registration subscription to the corresponding client group 102 requesting TM registration subscription, and registers and stores the entered subscription information in the TM information database 108e. . Preferably, when the client has already registered with another service of the site providing the search service in the present invention and wants to register TM, an input window for indicating agreement on a predetermined condition or an additional information input window required for TM registration corresponds. It transmits to the client and registers and stores the entered subscription information in the TM information database 108e. Here, the subscription information may include information such as personal information, identifier (ID), password, career, email address, mobile phone number, etc. of the client, and may include items that directly or indirectly reflect the level or quality. It is not limited to this.

In addition, the TM processing unit 106d has a user evaluation reliability indicating the reliability of the evaluation of the other TM client's TPs for each TM client 102a, a point indicating the accumulated contribution to the search service, ability, quality, and the like (or At least one of similar functions such as mileage, grade, air service, and the like, is stored and managed in the allocation area for the corresponding TM client 102a of the TM information database 108e.

Then, the hyperlink processing unit 106e receives the inter-TP hyperlink connection request signal and stores hyperlink information from the corresponding source TP requested for connection to the target TP in the hyperlink database 108g, and in the TP retrieval unit 106b. Receive a specific hyperlink selection (click) signal of a hyperlink list for a specific TP included in each search result to be generated, and then target TP in the TP information database 108e according to the hyperlink information stored in the hyperlink database 108g. The detection result is included in the new search result and transmitted to the search user client 102e which has transmitted the hyperlink selection signal. Preferably, before storing the hyperlink information requested for connection, the method may include requesting review and approval of the hyperlink connection request from the TM client 102a of the target TP, and obtaining the approval. New search results may include user add-ons.

Next, the evaluation processing unit 106f calculates the hyperlink popularity using a predetermined calculation function based on the hyperlinks between the TPs, and hyperlinks all the TPs stored in the TP information database 108e at the time of recalculation of the predetermined period. Re-populate popularity. Preferably, the predetermined calculation function uses at least one of the number of hyperlinks, the popularity of the source TP, and the frequency of using the hyperlinks.

Further, the evaluation processing unit 106f calculates the TP ranking value of each corresponding TP for determining the relative priority on the position according to the placement position of each TP summary item of the TP summary list (list) included in each search result. do. This is calculated using a predetermined TP ranking value calculation function, and is calculated according to input values corresponding to each element variable based on a formula that applies relative weights to the TP ranking value element variables. The input values of the TP ranking value element variables are periodically re-calculated and stored in the TP information database 108e at the time of collective retrieval and the corresponding values are detected when the TP ranking value calculation function is applied. Substituting the same or transformed value into the corresponding element variable of the predetermined TP ranking value calculation function, the input value of the real-time TP ranking value element variable detects the stored measured value of the element variable in the TP information database 108e or in real time. The measured value calculated as is or converted into the corresponding element variable of the TP ranking value calculation function, and the TP ranking value is calculated through this.

Then, the evaluation processing unit 106f receives the selection signal for any function among the user additional functions included in each search result from the search user client 102e, and counts the number of times of performing each selected function each time the corresponding function is performed. Increasing the value of the data, accumulating the distribution of user evaluation scores each time a user evaluation of the TP is performed, and storing it in the allocated storage space of the TP information database 108e; At least one process of increasing the variable representing a is performed, and the data is stored in the TP information database 108e.

In addition, the evaluation processing unit 106f checks the status data of the TPs registered by the TM client 102a or the search user client 102e that has performed user authentication processing (i.e., user login) for the TP information database 108e. And store and register with the search service server 106 the user evaluation popularity of each TP and the TP evaluation (that is, the evaluation signal is transmitted to the search service server 106) through each TM client 102a. After collectively recalculating and assigning the user evaluation reliability indicating the evaluation reliability of each TM (or TM ID), it is stored and managed in the TM information database 108f.

When the member authentication request signal is received from the TM client 102a, the authentication processing unit 106g transmits a login information input window for member authentication to the corresponding TM client 102a, and transmits the input authentication information (login information). If authenticated in comparison with the member information registered and stored in the TM information database 108f, the corresponding TM client 102a is allowed to log in to the search service server 106. Here, the login information (authentication information) is configured to include a user ID and password (password), in addition, it is possible to enter all or part of the social security number.

On the other hand, the control unit 106h includes the TP processing unit 106a, the TP retrieving unit 106b, the topic processing unit 106c, the TM processing unit 106d, the hyperlink processing unit 106e, the evaluation processing unit 106f, and the authentication processing unit 106g. Performs a function of controlling the operation and storing the corresponding information according to the operation of each operation in the storage device 108.

Next, the storage device 108 registers and stores the personal information for the membership entered in the search user client 102e and the related information of the pre-registered TM, the search service server 106. A topic information database 108d and a TP information database 108e which register and store information related to each of the topics and TPs registered in the TP are stored and managed. Preferably, an index structure database 108a for index word-topic mapping, an index structure database 108 b for topic-TP mapping, an index structure database 108 c for index word-TP mapping, and a hyperlink for storing hyperlink connection information between TPs. And at least one of the link databases 108g.

For each database management of the storage device 108, each database may be created and managed using Oracle, Sybase, MS-SQL, MySQL, and the like, which are common Database Manager Systems (DBMSs). In addition, for each index structure database, not only such a DBMS but also a search engine SF-1, K2, etc., or a database having its own file structure can of course be constructed. Here, when using a search engine, dynamic index, variable length reverse file list processing, fast partial search, morphological analysis, and complex noun decomposition based query processing can be utilized.

Next, the topic-based search service providing system having the above-described configuration receives a TM registration request signal from any client and registers it as a TM, creates a TP for a specific topic, registers and stores the created TP, A process of registering and storing a hyperlink connection of such a TP will be described.

2 is a flowchart illustrating a process of registering a TM according to a TM registration request signal according to the present invention, and FIG. 3 illustrates a process of registering a TP created after user authentication and registering a hyperlink connection according to the present invention. 4 is a flowchart illustrating a search service homepage according to the present invention. These drawings describe the TM registration and TP creation process.

2, 3, and 4, the search service server 106 checks whether a TM registration request signal is received when any search user client 102e sends a registration request signal (step 202). Here, when the TM registration menu provided to the search user client 102e connected to the search service server 106 is selected as shown in FIG. 4, a registration request signal is transmitted to the search service server 106.

When the TM registration request signal is received as a result of the check in the step 202, the TM processing unit 106d of the search service server 106 displays the TM member registration information input window via the wired / wireless communication network 104 in the corresponding search user client 102e. (Step 204).

Through this, the search user client 102e transmits a TM registration request signal together with the TM registration information input (step 206), and the TM processing unit 106d receives TM registration information and adds it to the TM information database 108f. The registration is saved (step 208). Preferably, before registering and storing in the TM information database 108f, the integrity of the TM registration information, whether or not the predetermined condition determined by the operator is satisfied, and the like are registered and stored. Here, TM registration information may include ID (ID), password, area of interest, education, major, occupation, specialty, email address, mobile phone number, etc., and various conditions (e.g. Of course, it may include the airborne score / rating indicating the degree of activity on any site.

In addition, the TM registration request signal not only directly supports TM registration, but also all kinds of indirect and detour subscription request signals that allow TM activity as a result even if a member of the corresponding service site or a member of another service site is subscribed. Of course it is included.

Accordingly, in the case of providing a search service that operates not only the search service according to the present invention but also other services (for example, a mini homepage, a blog, a cafe, etc.), the members who subscribe as members to other services are required for TM registration. Of course, it is also possible to receive and register only personal information (for example, agreement to terms and conditions).

On the other hand, the search service server 106 connects to the search service server 106a through the corresponding TM client 102a of the registered TM and performs the activity, and accesses and monitors the search service server 106a through the client. In operation 210, an ID of a specific mentor assigned to teach is designated and stored. Here, in the mentor field of the record related to the corresponding TM in the TM information database 108f, the ID (the ID ID or the operator ID) of the corresponding mentor designated as the mentor is registered and stored, and the mentor is searched if it is not specified. The service server 106 designates a mentor ID according to a predetermined criterion and registers and stores information on the mentor ID. As another method, when the TM registration information is input, the mentor information (for example, a pre-registered TM or an ID of a mentor) input together may be registered and stored in a mentor field of a record related to the TM.

On the other hand, while the TM registration information is already registered, the search service server 106 checks whether a user authentication request signal is received (step 302).

As a result of the check in the step 302, when a user authentication request signal is received, the authentication processing unit 106g of the search service server 106 performs a user information input window (for example, a login input) for performing user authentication processing. Window, etc.) to the corresponding TM client 102a, and user authentication is performed on the user authentication information entered through this input window to allow (log in) the search service server 106, if appropriate (step). 304). As an example, FIG. 4 is a diagram illustrating a TP search homepage according to the present invention, and may input and log in user authentication information (login information) through a TP search homepage including a login input window.

Then, the search service server 106 checks whether a TP creation request signal is received from the logged in TM client 102a (step 306).

As a result of the check in the step 306, when the TP creation request signal is received, the TP processing unit 106a transmits a TP creation tool to the corresponding TM client 102a for TP creation for any topic (step 308). . Through this, the TM client 102a creates a TP according to the edit (creation) command signal input through the input / output terminal (step 310). As an example, FIG. 5 is a diagram illustrating a TP editing tool web page according to the present invention. A TP for a specific topic may be created through the TP editing tool web page. Of course, one or more topics related to the contents of the TP may be designated as the specific topics.

Here, the TP editing tool can download from the search service server 106 in the TM client 102a, create a TP, and store it in the storage device of the TM client 102a, such as HTML, java script, asp, etc. You can add code, select and add edits to content and search results from other search sites, edit and modify TPs that are already registered, or request registration for new topics.

Next, the search service server 106 checks whether the registration request signal of the TP created from the TM client 102a is received (step 312). For example, when the detection target TP is selected through the storage TP view / detection menu as shown in FIG. 5, the target TP is loaded on a web page, and when the topic list view menu is selected, the registered topic list and the topic classification system are selected. After receiving some or all of them and displaying them on a web browser, a topic corresponding to the TP is designated, and a registration request signal for the TP is transmitted through the TP registration menu.

When the registration request signal of the TP created as a result of the check in step 312 is received, the TP processing unit 106a sends a message requesting review and approval of the contents of the TP through at least one of an SMS server, an MMS server, and an e-mail server. , At least one of the other TM clients 102a, the mentor client 102d, and the operator client 102f who have registered the TP in the registration designation target topic, to obtain an acknowledgment of approval (step 314). As an example, FIG. 6 is a diagram illustrating an e-mail regarding a TP registration review and reply request according to the present invention. FIG. 7 is a diagram illustrating an e-mail in the form of a questionnaire request form of a TP registration review request according to the present invention. 8 is a diagram illustrating a TP registration review request response status table according to the present invention, it is possible to perform the cumulative analysis and distribution of response distribution for the approval of registration of the TP created through the request email and response status table illustrated in these drawings. have.

According to the result of the response distribution analysis, the TP processing unit 106a assigns the approved TP to the index word-topic index structure database 108a in the storage device 108 and the index structure database 108b for topic-TP mapping for the designated topic. Or index into the index structure database 108c for index word-TP mapping and register and store in the TP information database 108e (step 316).

In addition, if no response arrives within the predetermined time period, it may be regarded as not opposed to the registration request and the TP registration requested as in step 316 may be performed.

On the other hand, in step 314, the review and approval request and acquisition procedure are omitted, and as in step 316, the TP is directly assigned to the index-topic index structure database 108a and topic-TP in the storage device 108 for the designated topic. It is of course possible to index into the index structure database 108b for mapping or the index structure database 108c for index word-TP mapping, and to register and store in the TP information database 108e. Preferably, the TP summary information of the TP is extracted and stored together with the corresponding TP ID in the TP information database 108e of the storage device 108 so that the TP summary corresponding to each TP can be detected.

In addition, it is preferable to check whether the client requesting the TP registration has obtained the authentication in step 312, and if the authentication is not obtained, a message requesting authentication is transmitted to the corresponding client, and the TP registration only after obtaining the authentication. Can be processed.

If the registration target topic of the TP is not specified, the TP processing unit 106a may recommend or designate a suitable topic. Here, the automatic recommendation or designation method may include, for example, a method for recommending a topic having a synonym list including keywords appearing in the title or body of the TP as a candidate, and the like. no.

As an example of the above steps, FIG. 9 is a diagram illustrating a relationship between an index word, a topic, a synonym list, and a TP according to the present invention, and FIG. 10 is an index word, a topic, a synonym list, and a TP according to the present invention. FIG. 11 is a diagram illustrating a detailed structure of a topic related information list according to an exemplary embodiment of the present invention. FIG. 11 is a diagram illustrating a detailed structure of a list of topic related information according to the present invention. Registration can be saved.

In more detail, in FIG. 9, the index word list accesses the index structure database 108a for index word-topic mapping or the TPs registered directly for the index word, which is a list of access information for all topics including each index word in the synonym list. An index structure database 108c for indexer-TP mapping, which is a list of information, is included. Here, the access information is a TP unique identifier (or identification number), and the unique identifier can detect the corresponding TP and TP summary information in the TP information database 108e, and this TP information database 108e is for example. It can be built using SQL server (DBMS).

In addition, FIG. 10 illustrates a relationship according to another embodiment of adding access information. In FIG. 11, the topic common TP includes information common to TPs registered in the same topic or common information related to a topic. ) Or by some collaborative editing work of TM clients 102a (e.g., in a wiki) or some automatic extraction and recording.

Here, the common information may include a topic definition, an additional description, a main attribute, and statistical distribution information of TPs registered in a corresponding topic (for example, an average input of TPs registered in a corresponding topic). Number of hyperlinks, etc.), meta information (e.g., the number of TPs registered on the topic, the list of newest registered / modified TPs, etc.), topic common posts, comments, topic identifiers, topic identification numbers, parent / child category topic identification Other semantic relations with other topics including names, at least one of a target topic identification name, and a synonym list may be included. Preferably, only one topic common TP can be registered for one topic, and the information of each TP is divided into sections to give a common part and a difference part in the form of distribution of registered TPs for each section. I can do it.

The search service server 106 checks whether a hyperlink connection registration request signal for the TP created from the TM client 102a is received (step 318). Here, the hyperlink connection request signal includes an identifier or access information (eg, url) of the source TP and an identifier or access information (eg, url, etc.) of the target TP. In addition, a target topic identification name (including a single index word without a synonym or a synonym group) in which the target TP is registered may be included, and a hyperlink name (name of a short description form of a conceptual relationship between two topics) may be included.

As an example, when the target TP search menu is selected as shown in FIG. 5, a web page for browsing the topics in which the target TP is registered and the TP registered in the corresponding topic is displayed. The hyperlink connection information for the target TP is displayed in the corresponding field. As shown in FIG. 12, the ID of the target TP, the corresponding topic identification name in which the target TP is registered, the ID of the source TP, the corresponding topic identification name in which the source TP is registered, the hyperlink name, etc. are designated through the hyperlink connection information. Transmits a hyperlink connection registration request signal.

When the hyperlink connection registration request signal is received as a result of the check in the step 318, the hyperlink processing unit 106e stores the hyperlink connection information of the corresponding TP to the target TP of the hyperlink database 108g of the storage device 108. The registration is stored in (step 320).

Here, before registering and storing the above-described hyperlink connection information, the hyperlink processing unit 106e sends a response request message regarding whether to review and approve the content of the hyperlink connection registration request to the TM client 102a of the target TP. It transmits using at least one of SMS and MMS, and registers and stores the hyperlink connection information when an acknowledgment response is received within a predetermined time or there is no response within a predetermined time. Here, the response request message includes the access information and the hyperlink name of the source TP, and preferably, further includes a target topic name of the target TP, thereby making it meaningful between the contents of the source TP and the topics of both TPs. Review the adequacy of the relationship. In addition, when receiving a response denied approval, the corresponding information can be transmitted to the TM client 102a of the source TP using e-mail, SMS, MMS, etc. Of course.

In addition to the hyperlink connection between the general TP and the TP as described above, in the hyperlink connection between the topic and the topic, or between the TP and the topic, the topic common TP representing the topic and containing common and essential information represents the hyperlink. Can be specified as the source or target of. That is, not only the general TP but also the topic can be a subject of the hyperlink connection, and if the TP is subdivided into sections, the connection with the specific section can be performed.

On the other hand, in the hyperlink connection, the hyperlink using the function to support the search target to recommend the upper, lower, and related topics by using a web page or a built-in thesaurus including a related word recommendation function, a topic search function, etc. It is desirable to provide a web page containing the link target topic to the TM client 102a.

Next, a process of registering a TP created for a specific topic in each database will be described with reference to FIGS. 9 and 10 as an example.

Referring to the method of registering the registration request TP based on the structure as shown in FIG. 9, the TP processing unit 106a of the search service server 106 stores the TP in the TP information database 108e and registers the stored TP. The access information for the newly registered TP is added to the topic structure index topic database 108b for mapping. Here, when registering a TP directly to a single index without a synonym group defined instead of a general topic having a synonym group, the access information for the newly registered TP is added only to the index structure database 108c for index-TP mapping.

In addition, not only simple words but also complex nouns (eg, college student missions, information retrieval, etc.), phrases (eg, home, blood, blood, notebook and desktop) , How to avoid the sun, etc.), sentences (eg, disappear with the wind, etc.) may be used. Here, if it is allowed to use more than one word, it is necessary to recognize them as a single expression that represents the same topic, regardless of whether they are spaced in the expression appearing in the search query. To this end, the topic name requested to be registered is used as it is in the index word list or topic node without modification, and the index word extracted by applying the stemming or index word extraction technique is also used. For example, WiseKMA4.0, which is an index word extractor of the commercial search engine SF-1 of 'WiseKnut (www.koreawisenut.co.kr)', or another morpheme analyzer and index word having the same function for the morphological analysis and index word extraction. You can also use an extractor.

For example, as shown in FIG. 10, the topic "Comparison of fuel consumption between Sorrento and Sonata", "Comparison of fuel efficiency between Sonata and Sorrento", "Comparison of fuel efficiency between Sonata and Sorento", "Comparison of fuel consumption between Sonata and Sorento", etc. It is necessary to match the various types of queries with the same similar meaning so that the topic can be detected.First of all, for the initial topic name "Sorrento vs. Sonata's fuel economy comparison", "Sonata vs. Sorrento's fuel economy comparison" is normalized form. Save to the topic node.

And, in order to be able to match similar pronouns (or variants) such as "Sorrento" and "Sorrento", "Sonata" and "Sonata", as shown in FIG. 10, index words of similar pronouns (synonyms) The access information (pointers or identifiers) from the entries on the list to the index structure database 108a for index word to topic mapping have the same value, so that these pseudophonic words are mapped to the same topics.

In addition, using topic morphemes and index word extractors (eg, WiseKMA 4.0) for topic distinguished names, search (i.e., i) and ending (i.e.) are separated, and compound nouns are decomposed, It extracts key words such as sonata, sorrento, and fuel economy. And, in the third topic index list of the topic nodes corresponding to each of the sonata (car), sorrento (car), and fuel economy comparison, which are topics reflecting the contextual meanings of the extracted index word sonata and sorrento topic, shown in FIG. As such, add access information (pointer) to the topic node "Comparison of fuel efficiency between Sorrento and Sonata" for the topic "Comparison of fuel efficiency between Sorrento and Sonata".

In addition, although the present invention has been described as an example for Korean, the application is not limited to Korean only, and the foreign language may be applied to a search service providing system by using a stemmer of a corresponding language and an index word extractor of the corresponding language. Of course.

Through the above-described process, the search query not only detects topics that exactly match the query string, but also changes the order of words, decomposition of compound nouns, spacing, etc., but the same query meanings the same topic. The topic can also be detected.

On the other hand, the processing method for the case of using the index structure for index word-topic mapping and the index structure for direct index word-TP mapping without using the index structure for topic-TP mapping is described in FIG. 9. The index registration method for the following topics will be described as an example.

First, extract the index words such as 'laptop', 'desktop', and 'performance comparison' for the topic using WISE KMA4.0, etc., and add the composite keyword topic “laptop and desktop” to index 1 of the complex index structure for each of them. Add each of the access information in "Performance Comparison of". The index1 structure functions as an index structure from element keywords for topics composed of compound keywords. Using this index structure, complex topics with similar meanings can be used for complex keyword queries with similar meanings, such as “Comparison of Laptop and Desktop Performance”, “Comparison of Laptop Desktop Performance”, and “Comparison of Desktop Notebook Performance”. TPs indexed at may be detected.

Next, as described above, a process of searching for a TP related to a topic matching a search query by a search user client in a state where a plurality of TPs for a plurality of designated topics are previously registered and stored in a topic-based search service providing system. Explain.

13 is a flowchart illustrating a process of searching for a TP created based on a topic according to the present invention.

Referring to FIG. 13, the TP search unit 106b checks whether a TP search request signal is received from the search user client 102e according to a search query input (step 1302). As an example, FIG. 14 illustrates a function of setting a TP search condition according to the present invention. The search user client 102e requests a TP search by setting a TP search condition and inputting a search query.

As a result of the check in step 1302, when the TP search request signal is received, the TP search unit 106b detects topics corresponding to (matched) the received search query (search term), and the TP registered in the detected topic. The first search result including the topic list (list) and the TP summary list is detected and transmitted to the search user client 102e via the wired / wireless network 104 (step 1304). . As an example, FIG. 15 is a diagram illustrating a first search result web page displaying a topic list, a TP summary, a TP detection, and a TP through a TP search according to the present invention. The topic list and the TP summary list may be displayed. .

Then, the TP retrieval unit 106b checks whether the selection signal for the specific topic in the topic list is received from the search user client 102e through the first search result (step 1306).

When the selection signal for a specific topic is received as a result of the check in step 1306, the TP search unit 106b detects a summary information list of TPs registered in the selected specific topic, and the topic list and the TP summary information list are displayed. After generating the included second search result, it is sent to search user client 102e (step 1308).

In addition, the TP retrieval unit 106b checks whether a selection signal of a specific TP summary item in the TP summary information list is received from the search user client 102e through the first search result or the second search result (step 1310).

When the selection signal of the specific TP summary item is received as a result of the check in step 1310, the TP corresponding to the specific TP summary information is included to generate a third search result and then transmitted to the search user client 102e ( Step 1312). Preferably, the third search result may include at least one of information of hyperlinks having the TP as a source and a selection menu related to a user additional function. As an example, FIG. 16 illustrates a third search result web page for detecting TPs and displaying TPs through a TP search according to the present invention, and displays a selection menu for detected TPs and user additional functions.

Meanwhile, in the present invention, the TP search unit 106b has been described as detecting a general topic having a group of synonyms matching the search query (search term). However, the TP search unit 106b detects a single index word that matches the search query and does not define a group of synonyms. Of course, it is also possible to generate and transmit a first search result including a summary list of indexed TPs in the detected and detected index words, in which case the steps 1306 and 1308 are omitted, and then step 1310 is performed. Done. The index word may be any one of a specific string, word, phrase, and sentence, and the matching method of the search query and the index word may be a complete match or a partial match matching method. In this case, the perfect match means that the search query and the index match exactly, and the partial match means extraction using the degree of similarity between the search query and the index. For example, morphological analysis can be used to separate functional words such as surveys and endings from search queries, and to use the similarity between extracted keywords and index words.

In addition, the search user client 102e may select an arbitrary user add-on through the third search result. The user add-on includes a recommendation function in which one search user recommends a TP to another search user, and a search user TP. At least one of user evaluation function, comment / reply function for TP content, scraping TP or some of TP content, mailing TP content to other search users, and outputting TP content. In addition, the blog may include a trackback function, an RSS function, and the like. Preferably, the method may include a function of rearranging and displaying the TP summary list included in the first search result and the second search result according to other criteria such as popularity, registration date, or the ranking value of each TP.

Here, the TP ranking value, which is a ranking value that determines the relative superiority of the placement of each TP on the TP summary list, includes the relationship between the TP and the search query, the TP related metadata, the popularity of the TP itself, the popularity of the TP creator, It is calculated from at least one of ranking value element types such as reliability. A detailed description of the TP ranking value calculation process will be described later.

17 is a flowchart illustrating a process of processing a hyperlink selection signal included in a TP search result according to the present invention.

Referring to FIG. 17, after transmitting each search result in the search process of FIG. 13 described above (step 1702), the search service server 106 receives a hyperlink selection (click) request signal from the search user client 102e. Check if it is received (step 1704).

When the hyperlink selection (click) signal is received as a result of the check in step 1704, the hyperlink processing unit 106e selects the target TP according to the hyperlink information stored in the hyperlink database 108g. Is detected in the new search result, and the hyperlink selection signal is transmitted to the corresponding search user client 102e (step 1706). As an example, FIG. 18 illustrates a table of information on a hyperlink-to-TP hyperlink according to the present invention, and detects a target TP for processing a hyperlink selection signal requested by a search user client 102e according to the information table. can do.

19 is a flowchart illustrating a process of performing processing of an optional user add-on selected through a TP search result according to the present invention.

Referring to FIG. 19, after transmitting each search result in the search process of FIG. 13 described above (step 1702), the search service server 106 receives an arbitrary user add-on request signal from the search user client 102e. Check if it is (step 1904). Here, the module for the user additional functions is stored in the control unit 106h of the search service server 106 so that the control unit 106h manages the performance of the user additional function.

As a result of the check in step 1904, when the user additional function selection signal is received, the control unit 106h of the search service server 106 performs the selected user additional function (step 1906). Here, when the control unit 106h executes a module corresponding to the user additional function, each corresponding module receives a necessary input from the corresponding client and transmits the processing result to the corresponding client. As an example, FIG. 20 is a diagram illustrating various types of user evaluation score display and input interface methods for a user evaluation function for a TP among user additional functions according to the present invention. You can enter In addition, FIG. 21 is a diagram illustrating a storage structure of user evaluation popularity-related data according to the present invention. The popularity according to the input user evaluation score may be stored in the TP information database 108e as such a data storage structure.

Thereafter, the search service server 106 detects a record of the TP database corresponding to the TP included in the search result through the evaluation processing unit 106f, and increments the execution frequency field value corresponding to the selected additional function. Save it again (step 1908). These number of execution field values are accumulated to contribute to the calculation of the popularity of the corresponding TP to be described later.

FIG. 22 is a flowchart illustrating a process of calculating hyperlink popularity and user evaluation reliability at the time of batch reproduction according to the present invention.

Referring to FIG. 22, the search service server 106 checks whether or not the batch recalculation time is reached according to a predetermined period (step 2202).

As a result of the check in step 2202, when it is time to collectively re-produce, the evaluation processing unit 106f calculates the hyperlink popularity using a predetermined calculation function based on the hyperlink between TPs. Hyperlink popularity is collectively recalculated for all TPs stored in the TP information database 108e (step 2204). Preferably, the predetermined calculation function reflects at least one of the number of hyperlinks, the popularity of the source TP, and the frequency of using the hyperlinks.

Next, the evaluation processing unit 106f recalculates the user evaluation popularity based on these data (step 2206).

Subsequently, the evaluation processing unit 106f checks the status data of the TPs registered by the TM client 102a or the search user client 102e that has performed the user authentication process (i.e., user login) for the TP information database 108e. Determining and assigning the user evaluation reliability which collectively indicates the user evaluation popularity of each TP and the evaluation reliability of each TM with respect to the TP created by another TM is stored in the TM information database 108f, and is managed. 2208).

Subsequently, the evaluation processing unit 106f collectively recalculates the measured value or input value of other batch recalculation TP ranking value element variables and stores the value of the corresponding element variable in the storage device 108 (step 2210).

Next, the process for recalculating the TP ranking value, hyperlink popularity, user rating popularity, and user rating reliability will be described in detail with respect to the above-described process.

<TP ranking value calculation>

The TP ranking value is calculated from at least one of ranking value element types, such as the relationship between the TP and the query, the metadata related to the TP, the popularity of the TP itself, and the popularity or reliability of the TP author.

The ranking value element of the TP-related metadata-related ranking value element type includes registration or modification time, the size of the TP file, the number of attachments / documents, etc., and the ranking value element of the popularity-related ranking value element type of the TP itself is TP. Direct visit count, comment / comment count, recommendation count, scrap count, print count, mailing count, user rating popularity score, hyperlink popularity score, etc. Factors include the TP writing TM's education level, TM's grade / mileage / points, occupation, and age.

On the other hand, since the ranking value element may have a variable value derived based on the measured value which is data accumulated through the measurement, it will be referred to as a ranking value element variable. For example, the number of TP direct visit hits (visitors or hits), which is a ranking value element variable, was visited by the TP directly through the selection (click) of a specific TP summary on the TP summary list on the search results page, rather than through a hyperlink. Has a measured value measured cumulatively.

In the present invention, a method for calculating a TP ranking value has a predetermined TP ranking value calculation function, and the TP ranking value calculation function includes input values and ranking value element variables derived based on measured values of each ranking value element variable. TP ranking value is calculated based on each weight. Here, the input value input to each element variable of the TP ranking value calculation function may be a measurement value per ranking value element itself or a converted value of the measurement value. An example of an input value obtained by converting a measured value is a ranking value element variable called registration or change time in a manner of giving '1 / (current time-registration or change time + 1)' to a variable called registration or change time. By converting the measured value of, the oldest TP that has been registered or changed has fewer inputs.

The TP ranking value is calculated by combining at least one or more of the ranking value element variable types in a predetermined manner, and the ranking value of any TP (a), RTP (a), is calculated by Equation 1 below.

RTP (q, a) = k1 * QR (q, a) + k2 * PTP (a) + k3 * PP (a) + k4 * MTP (a)

(0≤k1, k2, k3, k4≤1, k1 + k2 + k3 + k4 = 1)

MTP (a) = l1 * DP (a) + l2 * STP (a) + l3 * ATP (a)

(0≤l1, I2, l3 ≤1, l1 + l2 + l3 = 1)

PTP (a) = w1 * HLP (a) + w2 * UEP (a) + w3 * REP (a) + w4 * SCP (a) + w5 * RECP (a) + w6 * VP (a) + w7 * MP (a)

(w1 + w2 + w3 + w4 + w5 + w6 + w7 = 1, 0≤w1, w2, w3, w4, w5, w6, w7≤1)

Here, QR (q, a) is the relation between the TP (a) and the search query (q), and PTP (a) is the ranking value excluding the QR (q, a) part of the TP ranking value RTP (q, a). (I.e., a ranking value related only to the query itself and independent of the TP itself), PP (a) reflects the personal popularity of the TP creator TM (TM points, mileage, ratings, pledges, etc.), and MTP (a) is the metadata of the TP. The popularity calculated from the associated ranking value element variable, HLP (a) is the popularity having a functional relationship according to the hyperlink connection structure, the number, the number of clicks of the hyperlink, etc. to the target TP (a), and the UEP (a) Popularity that has a functional relationship to User Evaluation of the contents of a), REP (a) is a reply (Reply), comments (comments, footers, 20 comments, one or two lines, etc.) on the contents of the TP (a) It is called variously according to service provider) Popularity that has a functional relationship with the number of times of posting, RSS connection, trackback (remote comment) connection, etc. Popularity that has a functional relationship to the number of times scraping, storing, embedding, and printing of content in P (a), RECP (a) uses the contents of the TP (a) Popularity that has a functional relationship to the number of times a member is referred to a member, VP (a) is the popularity that has a functional relationship to the number of visitors (page views) of the TP (a), and MP (a) is the Popularity, which has a functional relationship with the number of times the mail containing the contents of the TP a is sent to the mail addresses of other members, DP (a) is the new date of the TP by subtracting the registration date of the TP (or the date of the last change). Is a measure of quantitative adequacy of a given TP, and ATP (a) is the attached content (image , Video, etc.) to measure the fidelity and diversity of TP content. Mihanda. Further, in Equation 1, w1, w2, w3,... … w7 represents the relative weight of each popularity factor item.

Meanwhile, input values of the ranking value element variable may be collectively recalculated at a predetermined predetermined period or may be calculated in real time during the TP search. In addition, the ranking value element variables for which input values are calculated in real time and the ranking value element variables for which input values are collectively recalculated at a predetermined predetermined period may be selectively designated by the operator client 102f or the search user client 102e. Can be.

First, the process of calculating the real-time ranking value element variable input value, which should calculate the measurement value of the TP ranking value element variable in real time, is described as an example. The value of the direct visit count per TP field in the TP information database 108e is described. It detects and sets it as the measured value and input value of the ranking value element variable called TP direct visit count (VP (a)).

In addition, the value of the accumulated count field is detected by increasing the value of the count field corresponding to the specific user additional function selected in the third search result and accumulating and storing the ranking. The measured value and the input value of the value element variable.

In addition, the measurement values and input values of other TP metadata and the popularity / confidence-related factor variables of the TM are shown through examples. 108e), and the value stored in the file size field of the record of the corresponding TP in the TP information database 243 when the ranking value element variable is the TP content size. Can be.

Secondly, the process of batch reproducing the input values of the TP ranking value element variables will be described. When the batch recalculation time is set, all of the TP ranking value element variables designated as all batch recalculation TP ranking value element variables are described. Recalculate all input values. Preferably, since the TP retrieval service must be performed in parallel even during batch retrieval, the batch retrieval is performed in a separate process or thread.

The user rating popularity reflects the cumulative distribution of user ratings for each TP among the ranking element variables belonging to the TP self-population type, and the hyperlink popularity is the user evaluation score for the semantic relevance of the hyperlink, It reflects at least one of the degree of popularity of other TPs providing the hyperlink link, the visitor inflow through the hyperlink (i.e., hyperlink clicks), the relative size of link utilization, and the number of inbound hyperlinks.

According to this method, the TP ranking value is calculated in real time at the time of retrieval of the measured values of the real-time calculated TP ranking value element variable and the collective recalculation TP ranking value element variable included in the TP ranking value calculation function or collectively re-established according to a predetermined period. The calculated value is calculated as it is or converted to the input value of the TP ranking value calculation function.

<Hyperlink popularity batch production>

At the time of batch recalculation, the hyperlink popularity of each TP is calculated through the hyperlink popularity calculation function of Equation 2 below, based on the cumulatively stored hyperlink selection (click) data (S421) (S421). .

HLP (a) = e1 * ∑ _b LC (b-> a) * PTP (b) * C (b-> a) / MC + e2 * ∑ _b PTP (b) * (1-k * (1-LL (b-> a))) + e3 * ∑ _b LN (b-> a)

(0≤e1, e2, e3≤1, e1 + e2 + e3 = 1)

Here, e1, e2, and e3 represent the relative weights of the respective terms. In an embodiment of the present invention, e1 = 1, but may have various other values. The first term in Equation 2 reflects the contribution of dynamic inflow through the hyperlink, the second term contributes to the weight or number of links, and the third reflects the number of inbound hyperlinks.

Also, PTP (b) is a ranking value of TP (b), and LC (b-> a) is a hyperlink click number of hyperlinks from TP (b) to TP (a) to all other TPs connected from TP (b). Divided by the number of link clicks, LL (b-> a) is the user rating score (score between 0 and 1 or between -1 and 1) for the semantic relevance adequacy of the hyperlink, and k is 0 or 1 In order not to make all the initial values of HLP (a) zero, the initial value of PTP (b) gives a very small nonzero value for all TPs in common. For example, 0.003 can be given.

C (b-> a) represents the number of hyperlink clicks from TP (b) to TP (a), and MC represents the average value of hyperlink clicks between any two TPs. In this case, C (b-> a) / MC will have more value if you provide more inflows than average inflows, and will have a smaller weight if you provide less inflows than average inflows.

In addition, if a hyperlink is not set from one TP (b) to another TP (a) but is set to a common TP of any other topic t, the following formula is applied to the topic from the TP (b). The increase in popularity due to the hyperlink to t is divided and given to all the TPs registered in the topic. The increase in popularity that contributes to any TP (a) in topic t is WC (b, a) = (LC (b-> t) * PTP (b) * C (b-> t) / MC) / n . Here, n is the number of target TPs to share, and the contribution mileage increase of TP (b) to the contribution of increasing popularity is WC (b, a) / (PTP '(a) + WC (b, a)) Shall be. PTP '(a) is the popularity value of PT (a) calculated before adding WC (b, a).

Referring to the process of calculating HLP (a) based on the above equations, for example, assuming that TP (a) is a TP having the TP identifier 14231 shown in FIG. Stored in linked list The value stored in the click frequency field of each of the source TPs (one TP identifier 10245 in FIG. 18) of the inbound hyperlink is detected. These values correspond to C (b-> a) of Equation 2. The current popularity value (ie, PTP (b) value) of the source TP (that is, the TP having a TP identifier 10245) is 0.75 extracted from the current popularity value field of FIG. 18.

Also, LC (b-> a) detects all the click counts of all outbound hyperlinks leaving C (b-> a) at TP (b) (i.e., TP identifier 10245) and adds them all together. Divided by 1, which is 1. MC represents the average click count per hyperlink divided by the number of hyperlinks by adding all the hyperlink click count values, and has 53.25 in FIG.

 Substituting these LC (b-> a), C (ba), MC, PTP (b) values into Equation 3, the new PTP (a) value is calculated, and the calculated value is added to the recent hyperlink popularity field. Save it.

In addition, in the above description, the hyperlink click may be applied as the number of times of use of the link by extending not only the hyperlink click itself but also a link type such as a trackback when a linking mechanism such as a blog is applied to the TP connection.

<Calculate the cost that the source TP provided the funnel to the target TP>

The contribution distribution ratio HLC (b, a) for the TP popularity due to the introduction of the inflow path through the hyperlink from the source TP (b) to the target TP (a) is defined by Equation 3 as follows.

HLC (b, a) = (LL (b-> a) * LC (b-> a) * PTP (b)) / (PTP (a))

Here, LL (b-> a) indicates a user rating popularity score for the hyperlink, and if the user rating for the hyperlink is not performed separately, the value is uniformly assigned to 1 and the popularity of the target TP (a) is improved. The mileage or points are given to the source TP (b) by reflecting the distribution ratio as contributed to HLC (b, a) itself, or HLC (b, a) * 3, or HLC (b, a) * PTP (a) and other various values can be freely selected according to the site operation characteristics.

In addition, the points of the TP (for example, mileage, grade, air hole, etc.) and the TP popularity are managed separately and derived by a separate calculation method. The TP popularity fluctuates over time with a range of values, but the point is the renewal portion of the TP popularity value and the cost of providing a hyperlink to another TP as the source TP, ie HLC (b, a) Add to the existing point value and accumulate.

However, TP popularity score PTP (b) is 0 when the value of granting points less than 0 is given from users due to poor or unhealthy contents of TP (b), and the relative weight of user rating popularity is large. It can be a smaller value. In addition, if it is determined that the content of TP b is unhealthy, PTP (b) is given a minimum value of less than zero as a penalty. In addition, in addition to the user evaluation of the contents of the TP, if the individual good and bad quality evaluations are also performed on the hyperlinks in the TP (that is, the information / significance of the hyperlink between the source TP (b) and the target TP (a) is inferior). Or when the user judges that the negative score is a lot of negative scores), the value of LL (b-> a) becomes smaller than 0 in Equation 3, resulting in HLC (b, a) being smaller than 0. Can be.

In this case, when PTP (b) is smaller than 0, the addition value becomes negative, so that the TP popularity point (or mileage, air hole, grade) is reduced. In such a case, the user's evaluation that the quality of the TP content is poor is dominant, and the overall page popularity becomes negative, and the point is reduced, so that the point content can be continuously reduced unless the page content is improved. In addition, since a value of a part of the TP popularity (that is, HLC (b, a)) is distributed to points to the source TP providing the inflow path, if a point decrease occurs due to the target TP (a) in the source TP (b), Actions may be taken to delete the link or automatically send a warning message to the TM client 102a of the target TP (a). For example, if you receive a point allocation that is reduced more than once, you can take steps to automatically delete the link.

<Population of User Rating Popularity and User Rating Reliability by TP>

Detailed description of the batch recalculation process of the user evaluation popularity and the user evaluation reliability corresponding to the ranking value element variable UEP (a) indicates that the TP created and registered by the TM client may be an unspecified number of search user clients 102e or other TM clients. Since the evaluation scores are variously given from the user, the user score grant status is accumulated and stored in the related data storage table, and the user evaluation popularity for each TP is calculated based on the accumulated stored data.

First, the user rating popularity for each TP is calculated based on the user evaluation score grant status data for each accumulated TP. This can be done in real time only for the target TPs detected in any search step or in batches for all TPs at the time of batch recalculation.

Based on the accumulated data, the user evaluation popularity of the TP is calculated using Equation 4 below.

UEP (i) = ∑ (UE _ij ) / N

Here, UE _ij is the jth user evaluation score for the TP (i) (for example, UEij ∈ {-5.-4-3, -2, -1,0,1,2,3,4,5) }), N denotes the number of times that various user evaluation scores are assigned to TP (i) over a period of time.

 For example, in FIG. 21, if the user evaluation popularity score grant status for the TP having the TP identifier 21032 is detected in the S-3 to S3 fields, and is substituted by Equation 4, the user evaluation popularity of the TP is calculated. ((-3x2) + (-2x3) + (-1x5) + (0x10) + (1x2) + (3x0)) / 23 = -0.52.

Meanwhile, the popularity of the user rating is reflected in the calculation of the TP ranking score, and according to the TP ranking value, the priority of the TP summary corresponding to the TP is placed on the TP summary list in the search result page. Done.

In addition, the method of reflecting the user's evaluation reliability is described. Instead of adding 1 uniformly to the frequency accumulation for the evaluation score for every user's TP evaluation, the evaluation reliability is added instead of or reflected in the evaluation reliability. Add the converted value. That is, when the search user client 102e or each of the TM clients 102a evaluates the TP, it is differentiated according to the user rating confidence given to that client's TM or search user (i.e., the ID of the TM or the search user's ID). This can reflect the evaluation score.

In addition, the user evaluation popularity UEP (i) of TP (i) reflecting the user evaluation reliability is calculated using Equation 5 below.

UEP (i) = ∑ (UE _ij * C _ij ) / ∑C _ij

Here, UE _ij is the j th user evaluation score for TP (i) (UE _ij ∈-5.-4-3, -2, -1,0,1,2,3,4,5}), C _ij denotes a cumulative user score of each client that has given an evaluation score UEij to TP (i).

The method of reflecting the user evaluation reliability in the user evaluation popularity of the TP will be described in detail. The search user client 102e or each logged-in TM client 102a having high user evaluation reliability and logged in to the search service server 106 is described. The TP, which is the sum of the evaluation scores of the users (or user IDs) of the users, indicates the evaluation scores given by the users of the search user client 102e who is low in user evaluation or not logged in or the TM client 102a who does not log in. It is intended to be higher than the sum.

And, the user evaluation reliability starts by giving the initial base score uniformly (e.g., 0.5, etc.) to both the newly registered TM client 102a or the user (i.e., user ID or TM ID) of the search user client 102e. Thus, the statistical distribution in the overall scoring status increases or decreases depending on whether they belong to the majority group or the minority group. In addition, a user of a search user client who is not logged in may uniformly assign a relatively low base score (eg, 0.0002) as a user evaluation reliability value.

As an example, referring to FIG. 23, the user evaluation score granting status for each TP will be described in detail. The scoring status is accumulated and stored for each registered user's user / TM or TM ID. A separate user-specific list of 102a or search user client 102e is maintained, and the corresponding score field stores a pointer to this list, respectively. For example, it is represented by evaluation score fields and other fields from the S-3 field to the S3 field of the TP information database 108e and the same scoring user list. The user / TM IDs of the TM client or search user client giving the same score to the same TP are included in the same list of the same scoring user list, with each component of the list being the TM of the corresponding TM in the TM information database 108f. Access information on the ID (eg, pointer, ID, etc.) is stored. This is represented by a pointer (or ID) from each list of the same scoring user list to the corresponding TP. The reason for maintaining the same scoring user list is that when the same score is assigned to the same TP, the same evaluation reliability score is given to this evaluation.

Next, referring to FIG. 23, the user evaluation signal processing process for the TP will be described in detail. The search service server 106 searches for the TP evaluation signal in which the TM having a TM ID of 3 is assigned an evaluation score of -1 to the TP ID 10245. When received from the user client 102e, the evaluation processing unit 106f detects the current user evaluation reliability score of the TM ID in the TM information database 108f and stores it in the cumulative score field of the S-1 field of the TP information database 108e. Add and resave, check if there is a pointer to the same scoring user list in the S-1 field of the TP information database 108e, follow the pointer and add the TM ID at the end of the list do. If there is no pointer, it creates a new list and stores a pointer to it.

Therefore, if the user evaluation score grant status for each TP is accumulated and stored, the user evaluation reliability can be collectively recalculated in the batch recalculation step.

Next, the process of calculating the user evaluation reliability will be described in detail. As shown in FIG. 23, a pointer to the same-score user list included in each field from the S-3 field to the S3 field for each TP is illustrated. Detects the same-score user list specified in the above, performs a predetermined evaluation of user evaluation reliability for each corresponding TM of the TM ID indicated by each TM ID of the detected list, and replaces the newly calculated value with the corresponding TM information database 102f. New store in user evaluation confidence field of TM. As an example, looking at the statistics of user ratings over a day for a TP, there were eight users who gave either -2 or -1 points, and the sum of these eight confidence ratings was 7.2, and either two or three points. If the number of trust ratings is 2.8 and the sum of the confidence ratings is 2.8, the person who gave -2 or -1 scores 7.2 / 10 = 72%, or 72%, in this assessment. Those who score -2 or -1 get only 28% confidence. The user evaluation reliability to be updated by the new evaluation is obtained by calculating the arithmetic mean of the user evaluation reliability obtained in the old and new evaluations. In other words, if a user who gave a -2 score had a previous evaluation reliability of 0.7 and a total of eight evaluations were accumulated, the new user evaluation reliability would be (0.7 * 8 + 0.72) / 9 = 0.70.

On the other hand, when the batch recalculation process is completed, the contents of the previously stored recalculation items are cleared and the new calculated values are prepared to be stored in the corresponding items again. However, when the number of user evaluations for a specific TP is very small, in order to prevent distortion due to the evaluation scores given by the malicious few users, the user evaluation reliability is not reflected when the number of user evaluations is less than a certain number of times. In addition, since fewer users participated in the evaluation imply less TP popularity, it is also desirable to reduce the score by a certain ratio (for example, 40%).

Next, a process of receiving a new topic registration request signal or a synonym addition registration request signal from an arbitrary client, verifying and receiving the content, and registering the new topic or synonym will be described.

24 is a flowchart illustrating a process of additionally registering a new topic according to the present invention, and FIG. 25 is a flowchart illustrating a process of additionally registering a topic synonym according to the present invention.

Referring to FIGS. 24 and 25, the search service server 106 may select any of the TM client 102a, the competing TM client 102b, the hyperlink TM client 102c, the mentor client 102d, and the operator client 102f. It is checked whether a registration request signal for a new topic or a synonym addition registration request signal for a specific topic is received from the client of the client (steps 2402 and 2502).

As a result of the check in the steps 2402 and 2502, if a registration request signal of a new topic or a synonym addition registration request signal for a specific topic is received, the topic processing unit 106c may use the mentor client 102d or the operator client 102f. Request at least one of the content verification (review) and approval (steps 2404 and 2504), and the topic processing unit 106c of the search service server 106 checks whether the approval is allowed (step 2404, 2504). Steps 2406, 2506).

As a result of the check in the steps 2406 and 2506, if approval is allowed, the topic processing unit 106c sends the requested new topic or a synonym for the specific topic to the topic information database 108d as a new topic or for a specific topic. And store it (steps 2408 and 2508). Here, the topic information is not duplicated by checking whether a new registration request topic or an additional request synonym for a specific topic is duplicated with a pre-registered topic or synonym without undergoing content verification, approval request, and approval approval. Of course, the database 108d may be added as a new topic or a synonym for a specific topic.

The above-described process will be described in detail with reference to FIGS. 26 and 27, when a new topic is registered, the TM client 102a, the competing TM client 102b, the hyperlink TM client 102c, and the mentor client 102d. ), Any of the operator clients 102f logs in to the search service server 106 and selects a new topic registration menu via a web page as shown in FIG. Accordingly, when a topic registration content is input and checked through a predetermined input window for inputting new topic registration related information, a new topic registration request signal is transmitted to the search service server 106 together with the topic registration content. In this case, the new topic related information may include elements including at least one of a topic identification name and a topic ID, a topic description, a simple descriptive description, a synonym list, and a higher category identification name.

When the search service server 106 receives the new topic registration request signal, the topic processing unit 106c adds and receives the received new registration request topic information to the topic information database 108d. Preferably, it is necessary to check redundancy whether a topic requested for new registration is already registered, and to be registered after reviewing and approving the validity of topic registration.

Here, the content verification (review) and approval process for topic registration will be described. The operator client 102f, the mentor client 102d, and other TM clients 102a who have registered TPs for the topics specified as the parent categories of the topic. E-mail containing a request for replying to a response within a predetermined response time, whether or not the topic requested to be registered through an e-mail server, an SMS server, an MMS server, etc. is registered, or whether the topic is registered. The Pros and Cons survey site sends at least one of the messages asking you to complete the pros and cons within the response timeframe.

Thereafter, when all of the response messages are received within a predetermined time for the approval request, or when the predetermined time has elapsed, response analysis for the approval is performed.

As a result of response analysis for this approval request, if the acceptance (or approval) is determined, the new registration request topic information is added to the topic node and the topic related information list of the topic information database 108d, and each of the topics included in the topic registration information is Additional synonyms are stored in the index structure database 108a for index term-topic mapping. Preferably, a message informing that the newly requested topic has been registered is transmitted to the corresponding TM client 102a through an email server, an SMS server, an MMS server, and the like. In addition, if a response is not received within a predetermined time, it can be regarded as approved and a new registration process can be performed.

Next, the process of adding a new synonym to the synonym list of a specific topic will be described in detail. When the user adds a synonym menu displayed on the search service homepage provided by logging in to the search service server 106, the synonym may be added. A predetermined input display window is provided for entering a target topic and synonyms to add. When at least one synonym to be added is selected by inputting a target topic and at least one synonym in the input window, the TM client 102a detects a synonym addition request signal including the input target topic and additional synonym information. Send to 106. Subsequently, the topic processing unit 106c of the search service server 106 that receives the synonym addition request signal detects a topic related information list of the target topic received together with the topic synonym addition request signal in the topic information database 108d. Additional synonyms are added to the synonym list included in the related information list and stored again in the topic information database 108d.

Preferably, additional registrations should be made only if the approval has been approved through deliberation of registration prior to additional registration of additional request synonyms in the synonym list of the corresponding topic. To this end, the operator client 102f, the competing TM client 102b, other TM clients who have registered TPs on the synonym registration target topic) as the subject of the content verification (review) and approval request, and the mentor client of the registration request TM client 102a. The content verification and approval status response message for the synonym addition is transmitted through the email server, the SMS server, the MMS server, or the like at least one of the 102d. For example, an email as shown in FIGS. 26 and 27 may be sent to transmit a content check and approval request message for an additional request synonym. Here, in the request for approval (deliberation), a minimum message indicating that there was a request for additional registration of synonyms, a message for requesting review because there is an additional registration request, a target topic name, and a message for requesting approval after deliberation with the additional target synonyms Etc. may be included.

In addition, the approval request message includes the access information for the topic. The reason for including access information for the topic is that the operator client 102f, the competing TM client 102b, and the mentor clients 102d of the registration request TM client 102a can easily view the various contents already registered on the topic. The access information is, for example, a web page URL for viewing the contents of the topic or a topic, a topic definition, a list of synonyms registered in the topic, a homepage of a service site, and a sub. It may consist of a temporarily assigned identifier in the form accessible from the page, but is not limited thereto, and other modified forms of access information are possible.

Preferably, taking into account the convenience of each client performing a content review, click on the URL of the page where the topic information is available directly in the email, and also review the suitability of the additional requested synonyms included with it, and further registration. It is possible to send an e-mail in the form shown in Fig. 27 in which the decision of approval or not can be immediately indicated and returned.

On the other hand, for the additional method after approval, the additionally requested additional synonym candidates are temporarily stored in a predetermined temporary registration store in the storage device 108, or once stored in the topic information database 108d and displayed in a "temporary registration" state. At least one scheme may be employed to temporarily store additional requested additional synonym candidates.

In addition, if no message for a review and approval response request is received within the prescribed time limit, it may be regarded as having no objection, and the registration request may be processed, and if there is at least one opposing message in the response message, It is considered that the content that needs to be reviewed is included, and the rejection message is sent to the additional registration request TM client 102a through at least one of the email server, the SMS server, and the MMS server together with the reason for rejection included in the response message. In addition, the TM client 102a that sent the additional registration request signal may send a message to the operator client 102f requesting a decision through appeal and review to convey an objection to the reject decision. This additional registration request signal can be received from the TM client 102a as well as from the logged in search user client 102e.

Next, any one of a registration request signal, a search request signal for a search query, a new topic registration request signal, and a synonym addition request signal for a specific topic are described. If a is received will be described for the corresponding process.

28A and 28B are flowcharts illustrating a process of providing a search service based on a topic according to the present invention.

Referring to FIGS. 28A and 28B, the search service server 106 may include a TM client 102a, a competitive TM client 102b, a hyperlink TM client 102c, a mentor client 102d, a search user client 102e, Registration request signal for any topic of the TP created via the wired / wireless network 104 from any one of the operator clients 102f, search request signal for search query, new topic registration request signal, synonym addition request signal for a specific topic Check which one is received (step 2802).

As a result of the check in step 2802, when the registration request signal of the created TP is received (step 2804), the TP processing unit 106a sends a message requesting review and approval of the contents of the TP to the SMS server, the MMS server, and the e-mail server. The response is sent to at least one of the mentor client 102d, the other TM clients 102a who have registered the TP to the registration target topic, and the operator client 102f through at least one of them (step 2806). .

The TP processing unit 106a then assigns the TP that has obtained the approval to the index language-topic index structure database 108a in the storage device 108 and the index structure database 108b for the topic-TP mapping or the index word-TP mapping for the designated topic. The index is indexed into the index structure database 108c and stored in the TP information database 108e (step 2808). As an example, FIG. 9 illustrates a relationship between an index word, a topic, a synonym list, and a TP according to an embodiment of the present invention, and FIG. 10 illustrates another embodiment between an index word, a topic, a synonym list, and a TP according to the present invention. FIG. 11 illustrates a detailed structure of a topic related information list according to the present invention. FIG. 11 is a diagram illustrating a detailed structure of a topic related information list. have.

Meanwhile, the search service server 106 checks whether a hyperlink connection registration request signal for the TP created from the client is received (step 2810). Here, the hyperlink connection request signal includes an identifier or access information (eg, url) of the source TP and an identifier or access information (eg, url, etc.) of the target TP. In addition, a target topic identification name (including a single index word without a synonym or a synonym group) in which the target TP is registered may be included, and a hyperlink name (name of a short description form of a conceptual relationship between two topics) may be included.

In addition, in step 2806, the review and approval request and acquisition procedure are omitted, and the index-topic index structure database 108a and topic-TP in the storage device 108 for the specified topic are directly designated as in step 2808. It is of course possible to index into the index structure database 108b for mapping or the index structure database 108c for index word-TP mapping, and to register and store in the TP information database 108e.

When the hyperlink connection registration request signal is received as a result of the check in step 2810, the hyperlink processing unit 106e stores the hyperlink connection information of the corresponding TP to the target TP of the hyperlink database 108g of the storage device 108. Register and save (step 2812). Here, as described with reference to FIG. 3, a user authentication process, a process of transmitting a TP editing tool, a process of designating a registered target topic of a created TP, a review and approval process of a hyperlink connection registration request, and the like may be included. Of course.

On the other hand, when the search request signal for the search query is received as a result of the check in step 2802 (step 2814), the TP search unit 106b detects topics corresponding to (matched) the received search query (search term). After detecting the summary list (list) of TPs registered in the detected topics to generate a first search result including the topic list (list) and the TP summary list, the search user client 102e through the wired / wireless network 104. (Step 2816). As an example, FIG. 15 is a diagram illustrating a first search result web page displaying a topic list, a TP summary, a TP detection, and a TP through a TP search according to the present invention. The topic list and the TP summary list may be displayed. .

Then, the TP retrieval unit 106b checks whether the selection signal for the specific topic in the topic list is received from the search user client 102e through the first search result (step 2818).

When the selection signal for a specific topic is received as a result of the check in step 2818, the TP search unit 106b detects a summary information list of TPs registered in the selected specific topic, and the topic list and the TP summary information list are displayed. After generating the included second search result, it is transmitted to the search user client 102e (step 2820).

Meanwhile, the TP retrieval unit 106b checks whether a selection signal of a specific TP summary item in the TP summary information list is received from the search user client 102e through the first search result or the second search result (step 2822).

When a check signal of the specific TP summary item is received as a result of the check in step 2822, a TP corresponding to the specific TP summary information is included to generate a third search result and then transmitted to the search user client 102e ( Step 2824). As an example, FIG. 16 illustrates a third search result web page for detecting TPs and displaying TPs through a TP search according to the present invention, and displays a selection menu for detected TPs and user additional functions.

Meanwhile, in the present invention, the TP search unit 106b has been described as detecting a general topic having a group of synonyms matching the search query (search term). However, the TP search unit 106b detects a single index word that matches the search query and does not define a group of synonyms. Of course, it is also possible to generate and transmit a first search result including a summary list of indexed TPs in the detected and detected index words, in which case the steps 2818 and 2820 are omitted, and then step 2822 is performed. Done. The index word may be any one of a specific string, word, phrase, and sentence, and the matching method of the search query and the index word may be a complete match or a partial match matching method. In this case, the perfect match means that the search query and the index match exactly, and the partial match means extraction using the degree of similarity between the search query and the index. For example, morphological analysis can be used to separate functional words such as surveys and endings from search queries, and to use the similarity between extracted keywords and index words.

In addition, the search user client 102e may select an arbitrary user add-on through the third search result. The user add-on includes a recommendation function in which one search user recommends a TP to another search user, and a search user TP. At least one of user evaluation function, comment / reply function for TP content, scraping TP or some of TP content, mailing TP content to other search users, and outputting TP content. In addition, the blog may include a trackback function, an RSS function, and the like. Preferably, the method may include a function of rearranging and displaying the TP summary list included in the first search result and the second search result according to other criteria such as popularity, registration date, or the ranking value of each TP.

Here, the TP ranking value, which is a ranking value that determines the relative superiority of the placement of each TP on the TP summary list, includes the relationship between the TP and the search query, the TP related metadata, the popularity of the TP itself, the popularity of the TP creator, It is calculated from at least one of ranking value element types such as reliability. The process of calculating the TP ranking value is the same as described above, and the process may include calculating the hyperlink popularity and the user evaluation reliability according to the batch reproduction time as described with reference to FIG. 22.

In addition, in the above-described TP retrieval process, as described with reference to FIG. 17, a process according to a hyperlink selection signal or a process according to a user additional function selection signal as described with reference to FIG. 19 may be performed. .

On the other hand, when the check result in step 2802, a registration request signal of a new topic or a synonym additional registration request signal for a specific topic is received (steps 2826 and 2834), the topic processing unit 106c receives a mentor client 102d. Request content verification (review) and approval from at least one of the operator client 102f, another TM client 102a that has registered a TP for a topic designated as a parent category of that topic (steps 2828 and 2836), and In response to the content verification and approval request, the topic processing unit 106c of the search service server 106 checks whether approval is allowed (steps 2830 and 2838).

As a result of the check in the steps 2830 and 2838, if the approval is permitted, the topic processing unit 106c sends a synonym for the requested new topic or a specific topic to the topic information database 108d as a new topic or for a specific topic. The registration is then stored (steps 2832 and 2840). Here, the topic information is not duplicated by checking whether a new registration request topic or an additional request synonym for a specific topic is duplicated with a pre-registered topic or synonym without undergoing content verification, approval request, and approval approval. Of course, the database 108d may be added as a new topic or a synonym for a specific topic.

Here, when the search service server 106 receives the new topic registration request signal, the topic processing unit 106c adds the received new registration request topic information to the topic information database 108d and stores it. Preferably, it is necessary to check redundancy whether a topic requested for new registration is already registered, and to be registered after reviewing and approving the validity of topic registration.

Therefore, by constructing a topic-based search service providing system, a TP for any topic is created and registered, a corresponding topic or a corresponding TP corresponding to an arbitrary search query is searched, and an optional user additional function is performed according to the search result. And register new topics or add synonyms to specific topics.

Meanwhile, as described with reference to FIGS. 28A and 28B, a registration request signal for any topic of a TP created through a topic-based search service providing system, a search request signal for a search query, a new topic registration request signal, and a specific topic When any one of the synonym addition request signals for is received, the process is performed separately for each case, but the corresponding process is performed in parallel with the corresponding process according to the search request signal for the search query. If any one of a registration request signal for a topic of the TP, a new topic registration request signal, and a synonym addition request signal for a specific topic is received, the process may be performed.

In the foregoing description, the present invention has been described with reference to preferred embodiments, but the present invention is not necessarily limited thereto. Those skilled in the art will appreciate that the present invention may be modified without departing from the spirit of the present invention. It will be readily appreciated that branch substitutions, modifications and variations are possible.

As described above, the present invention relates to a search service system and a method for searching and providing user registration contents related to a search query inputted by a client and a search service server connected to a wired / wireless communication network. Unlike the conventional method that does not provide an identification method for homonyms and synonyms, TM registration request signal, TP registration request signal that is content for a specific topic, registration request signal for a new topic, new synonym addition request signal for a specific topic, A client transmitting one of a hyperlink connection request signal for another TP, a search request signal for a search query, and a user add-on selection signal to a search service server through a wired / wireless network, and performs signal processing for each request signal. And search based on the search request signal for the search query. By constructing a topic-based search service providing system including a topic list related to a query, a summary list of TPs registered in a topic selected in a client, and a search service server for transmitting any one of the TPs to a client through a wired / wireless communication network, By constructing a topic-based search service providing system that can identify homonyms and synonyms, we can effectively provide search results that match the meaning of the user's search query.

In addition, in the topic-based search service providing system, the creation and registration of topic-based TPs, the provision of TPs for specific topics, and the provision of hyperlinks to other related TPs provide relevant information according to the intimate topic association between registered topics. Improve access to

In the topic-based search service providing system, input values of hyperlink popularity, user evaluation popularity, user evaluation reliability, and other ranking value factor variables between TPs are calculated and reflected in ranking values for each TP, and evaluation reliability of arbitrary TPs is calculated. And the quality of the contents and search of the TP can be improved.

In addition, through the topic-based search service provision system, it is possible to systematically cope with the problem of query query and synonym processing, and the community network according to the provision of compensation and regulation through the development of guidance, cooperation, competition among each client, etc. The setting can effectively maintain the quality of the search service.

In addition, according to the hyperlink connection in the topic-based search service providing system, the search quality is directly or indirectly evaluated and the less relevant TP is eliminated, and the more relevant or popular TP is strengthened, so the explicit meaning of the related topic is enhanced. Can form an enemy network.

In addition, by registering information related to a single topic or topic in a search service providing system and promptly verifying the reliability of the information, it is possible to promptly promote new information related to the relevant topic or topic, and have a high quality related to the topic. By providing an explicit gathering point for gathering information and information from a variety of sources, it is possible to improve the quantitative abundance and integrated access of information on relevant topics.

Lastly, by constructing a topic-based search service providing system, it is possible to provide a semantic-based search service that can obtain synonyms for various search queries including related words, dictionary of related words, etc. at low cost.

Claims (54)

A system including a storage device and a client and a search service server connected to a wired / wireless network to provide a search service for an arbitrary search query. A Topic Manager (TM) client requesting registration of a Topic Page (TP) created for any topic, a search user client requesting a search of the search query, and another created on the same topic as the specific topic of the created TP Providing a request and response for the search service, including a competing TM client requesting registration of a TP, and requesting new synonym registration for a new topic or specific topic, TP processing unit for processing the registration request for the TP registration request from the client, TP search unit for providing any one of the topic, TP, TP summary list according to the search request from the client, and controls each process A control unit for storing corresponding information in the storage device, a topic processing unit for processing a registration request according to a registration request for a new topic or a new synonym for a specific topic, and a TM from the client. A search from the client including a TM processing unit for processing a registration request according to a registration request and an evaluation processing unit for calculating a TP ranking value of each TP for determining a relative priority on the placement position of TM summary items from the client The response result related to the service request is transmitted through the wired / wireless network. The search service server for transmitting to the client Topic-based search service providing system comprising a. delete The method of claim 1, The client, A hyperlink TM client for creating another TP hyperlinked to the created TP and transmitting a registration request signal thereto; A mentor client managing the TM client or the competing TM client, Operator client operating the search service providing system Topic-based search service providing system further comprising. The method of claim 1, The client, TP (Topic Page) creation or support request signal, created TP registration request signal, hyperlink connection request signal, user add-on request signal, new topic add registration request signal, additional registration request signal of information about a specific topic, user add-on The topic-based search service providing system, characterized in that for transmitting any one of the selection signal to the search service server through the wired or wireless communication network. The method of claim 1, The client, A topic-based search service providing system according to claim 1, wherein any one of the comments, posts, and trackbacks of the TP is written, and an additional request signal for the TP is transmitted to the search service server through the wired / wireless communication network. The method of claim 1, The TP search unit receives a search request signal for a search query from the search user client and lists a corresponding topic, a TP summary list registered in a unique topic corresponding to the search query, and a topic selected from the list of topics. And a TP summary list registered in the TP summary list and a TP selected from the TP summary list to the search user client. delete The method of claim 1, The search service server, A hyperlink processor for receiving hyperlink connection request signals from the TM client, the competing TM client, the mentor client, or the operator client to process hyperlink connection information of a TP for a specific topic and a TP for another topic; An authentication processor for receiving an authentication request signal for logging in to the search service server from the TM client, the contention TM client, the mentor client, or the operator client and processing an authentication request thereto. Topic-based search service providing system further comprising. The method of claim 1, The TP ranking value is based on at least one or more ranking value element types of the relation between the TP and the search query, the TP related metadata, the TP itself popularity, the popularity of the TP author, and the reliability of the TP author. Topic-based search service providing system, characterized in that calculated. The method of claim 1, The evaluation processing unit, the topic-based search service providing system, characterized in that the re-calculation of the input value or measured value of the batch reproduction TP ranking value element variable including the hyperlink popularity and user popularity of each TP according to a predetermined period . The method of claim 1, The control unit may further perform processing operation control of the topic processing unit, the TM processing unit, the hyperlink processing unit, the evaluation processing unit, and the authentication processing unit, and control to register and store information related to the processing in the storage device. Topic-based search service providing system. The method of claim 1, The search service server transmits a TP creation tool according to the TP creation support request signal, registers with the storage device based on a registration target designation topic according to the created TP registration request signal, and the TM registration request signal, The topic-based search, wherein the corresponding information is registered in the storage device according to at least one of a hyperlink connection request signal, the new topic addition registration request signal, and a registration request signal for adding or modifying information on the specific topic. Service delivery system. The method of claim 12, And the search service server performs a user authentication process before registering with the storage device. The method of claim 1, The storage device, An index structure database for index word-topic mapping that registers and stores index word-topic mapping information provided from the search service server; Index structure database for topic-TP mapping that registers and stores the topic-TP mapping information provided from the search service server. Topic-based search service providing system comprising a. The method of claim 1, The storage device, Index structure database for index word-TP mapping that registers and stores index word-TP mapping information provided from the search service server. Topic-based search service providing system comprising a. The method according to claim 14 or 15, The storage device, A topic information database for registering and storing information on any topic provided from the search service server; A TP information database for registering and storing TP information for any topic provided from the search service server; TM information database for registering and storing TM information provided from said search service server Topic-based search service providing system further comprising. The method according to claim 14 or 15, The storage device further comprises a hyperlink database for registering and storing hyperlink information between any TP and another TP provided from the search service server. The method according to any one of claims 1, 3 to 6 or 8 to 15, The topic is a topic-based search service providing system, characterized in that the group of search words including a search query having a form of any one of a specific string, phrase, clause, sentence. The method of claim 18, The topic may include at least one of a topic identification name, a descriptive description of the topic, a higher category topic identification name, and a synonym list. The method according to claim 1, 3 or 8, The TM client is a topic-based client, which transmits a TM registration request signal for a corresponding user to the search service server, creates a TP for an arbitrary topic, and transmits a registration request signal to the search service server. Search service provision system. The method according to any one of claims 1, 3 to 6 or 8 to 15, The TP is a topic-based search service providing system, characterized in that the post form of a web page or blog containing at least one of information, content, edits of other search engine search results, and related programs for a specific topic. A method comprising a storage device and a client and a search service server connected through a wired or wireless communication network to provide a search service for an arbitrary search query. Checking whether a search request signal for the search query is received from the client; When the search request signal is received, the first search result is generated by including a unique topic or a topic list corresponding to the search query and a TP summary list registered in correspondence with the specific topic or the unique topic. Transmitting to the search user client through; Checking whether a selection signal of a specific topic is received through the first search result when there are two or more topics in the topic list; Generating a second search result including a list of TP summaries registered corresponding to the topic when the selection signal of the specific topic is received, and transmitting the second search result to the search user client through the wired / wireless communication network; Checking whether a summary selection signal of a specific TP is received through the first search result or the second search result; When a summary selection signal of the specific TP is received, at least one of a TP corresponding to the summary of the TP, a hyperlink related information list using the TP as a source TP, a selection menu related to a user additional function including user evaluation and user recommendation Generating a third search result including one and transmitting the third search result to the search user client through the wired / wireless communication network; During the above steps, checking whether a new topic registration request signal or a synonym addition request signal for a specific topic is received from the client; Registering and storing information about the new topic registration request signal or a synonym addition request signal for a specific topic when the new topic registration request signal or the synonym is not duplicated with a previously registered topic or synonym; During the above processes, when a predetermined period is reached, a process of calculating a measured value or input value of a batch recalculation TP ranking value element variable for each TP and a user evaluation reliability Topic-based search service providing method comprising a. The method of claim 22, The providing method, Checking whether a registration request signal for any topic of the TP created from the client is received; Registering the created TP on the storage device based on a registration target designation topic when a registration request signal of the created TP is received; Topic-based search service providing method further comprising. The method of claim 23, The providing method, Checking whether a client receives a hyperlink connection registration request signal from the client to another TP for the created TP; Registering hyperlink connection information with the storage device when the hyperlink connection registration request signal is received; Topic-based search service providing method further comprising. The method of claim 23, The providing method may further include performing a user authentication process before registering with the storage device. The method of claim 23, The providing method may further include a step of obtaining content verification and approval from any one of another TM client, a mentor client, and an operator client who registered the TP in a registration target topic when a registration request signal of the created TP is received. How to provide search service based on. The method of claim 23, The providing method may further include transmitting a TP creation tool to the client through the wired / wireless communication network when the TP creation request signal is received, and transmitting the TP created through the transmitted TP creation tool to the search service server. Topic-based search service providing method comprising. The method of claim 23, The providing method may further include receiving a TM registration request signal and registering TM registration information with the storage device. The method of claim 23, The providing method may further include specifying a mentor for the registered TM. The method of claim 23, wherein The providing method is a topic-based search service providing method, characterized in that at least one topic to be registered for the TP is specified, and the search service server is recommended or specified if not specified in the client. The method of claim 22, The providing method may further include receiving a connection request signal including hyperlink connection information from a TP to another TP for the specific topic, and performing hyperlink connection to the corresponding TP. . The method of claim 31, wherein The hyperlink connection is a topic-based search service providing method characterized in that the connection between the topic and the TP or the topic. The method of claim 32, When the target is a topic, the hyperlink connection includes topic common information instead of the general TP registered to the topic, and provides a topic-based search service, which connects a topic common TP representing the topic. Way. The method according to any one of claims 31 to 33, The hyperlink connection includes at least one of an identifier of a target TP, a corresponding topic identification name in which the target TP is registered, an identifier of the source TP, a corresponding topic identification name in which the source TP is registered, and a hyperlink name as the information. Topic-based search service providing method. delete The method of claim 22, The user additional function may include at least one of the user evaluation, the user recommendation, comment, scrap, mail transmission, output, trackback, and RSS. The method of claim 22, The first search result and the second search result are displayed by sorting the TP summaries of the TP summary list according to one of the creation date of the TP, the popularity criterion of the TP, and the ranking value of the TP. Topic-based search service. The method of claim 37, wherein The TP ranking value may include at least one ranking value element type among the relation between the TP and the search query, the TP related metadata, the popularity of the TP itself, the popularity of the TP creator, and the reliability of the TP creator. Topic-based search service providing method characterized by using. The method of claim 38, The TP-related metadata-related ranking value element type may include at least one of a registration or modification date, a size of a TP file, and the number of attachments / documents. The method of claim 38, The type of the ranking-related ranking value element of the TP itself may include at least one of a TP direct visit count, a comment count, a recommendation count, a scrap count, an output count, a mail count, a user rating popularity score, and a hyperlink popularity score. Topic-based search service providing method characterized in that. The method of claim 38, The popularity of the TP writer or the reliability-related ranking value element type of the TP writer may include at least one of the educational attainment of the TP authoring TM, the grade / mileage / point of the TM, the occupation, and the age. How we deliver the service. The method of claim 22, The providing method, Checking whether it is time for batch recalculation according to the predetermined period in the step of calculating the user evaluation reliability; Recalculating the measured value or input value of the batch reproduction TP ranking value element variable including the hyperlink popularity per TP or the user evaluation popularity per TP when the batch reproduction time is reached; Calculating the user evaluation reliability according to the cumulative evaluation of the TP and recalculating the hyperlink popularity or the user evaluation popularity based on the user evaluation reliability; Topic-based search service providing method comprising a. The method of claim 42, And wherein the hyperlink popularity is calculated based on at least one of cumulatively stored hyperlink selection data, the number of inbound hyperlinks, and user rating popularity of the source TP. The method of claim 43, The hyperlink popularity is a topic-based search service providing method, characterized in that reflecting the number of inflow paths provided for a predetermined period via a hyperlink from the source TP to the target TP. The method of claim 43, The hyperlink popularity is a topic-based search service providing method, characterized in that a reward of a mileage or point scheme of a predetermined rate in return for the TM of the source TP providing the inflow path. The method of claim 43, The user evaluation popularity for each TP is a topic-based search service providing method, characterized in that calculated based on the distribution of the user evaluation score data for the accumulated TP. The method of claim 43, And the user evaluation reliability is calculated based on a distribution of the user evaluation score data for the accumulated and stored arbitrary TPs. delete The method of claim 22, The providing method, Requesting verification and approval of contents of a new topic or additional synonym when the new topic registration request signal or a synonym addition request signal for a specific topic is received; Checking whether approval of the topic or additional synonym is permitted in accordance with the requested verification and approval; Registering and storing information about the topic or additional synonym in the storage device when the approval of the corresponding topic or additional synonym is permitted or if no response is received within a predetermined time period Topic-based search service providing method further comprising. The method of claim 22, 23 or 49, The topic is a topic-based search service providing method, characterized in that the group of search words including a search query having a form of any one of a specific string, phrase, clause, sentence or a synonym for the topic. 51. The method of claim 50, The topic may include at least one of a topic identification name, a descriptive description of the topic, a higher category topic identification name, and a synonym list. The method of claim 22, 23 or 49, The TP is a topic-based search service providing method, characterized in that the post form of a web page or blog including at least one of information, content, edits of other search engine search results, related programs for a specific topic. The method of claim 52, wherein The TP further includes one topic common TP representing the topic in providing a hyperlink connection and topic common information between the topic and another topic or between the TP of the topic and another topic for any topic. Topic-based search service providing method. The method of claim 53 wherein The topic common TP includes information content or meta information related to a topic or information commonly included in TPs registered in the same topic, and common information includes a definition of a topic included in TPs registered in the same topic. Description, key attributes, statistical distribution information of the TPs registered for that topic, topic common posts, comments, topic distinguished names, topic identification numbers, other semantic relationships with other topics including parent / child category topic identifiers, and At least one or more of a target topic identification name and a synonym list may be included, and as the meta information, at least one of the number of TPs registered in the corresponding topic and a list of the latest registration / change TPs is provided. Way.

Images