KR101111086B1 - Agent-Based Personalized Multimedia Jukebox System - Google Patents

Abstract

The present invention relates to an agent-based personal multimedia jukebox system and a multimedia recommendation method.
According to an aspect of the present invention, there is provided a method for recommending multimedia content, wherein the terminal is executed in a terminal having at least a function of searching for multimedia content and storing the multimedia content by accessing a server providing the multimedia content. A first step of acquiring and accumulating weblog information in a process of searching for selecting multimedia contents by accessing the server; A second step of determining, by the terminal, a behavior type of the user, wherein the behavior type is classified into a predetermined number of types by the terminal based on the weblog information accumulated in the first step; And a third step of the terminal recommending the multimedia content based on the behavior type of the user determined in the second step, from among the plurality of multimedia contents recommendable to the user.

Description

Agent-Based Personalized Multimedia Jukebox System

The present invention relates to an agent-based personal multimedia jukebox system and a multimedia recommendation method, and more particularly, to select recommended content from among a plurality of contents stored in the personal multimedia jukebox. The present invention relates to an agent-based personal multimedia jukebox system and a multimedia recommendation method for recommending content according to an individual's behavior type and preference genre by classifying by using the data, and analyzing personal preference genres using past usage history.

Recently, due to the development of Internet technology, a lot of information exists, and many people share information through the Internet. In particular, as multimedia contents (music, videos, photos, etc.) have been digitized, users can easily download and enjoy desired contents through the Internet.

In addition, as the performance of the computer increases, the storage space increases, and a large amount of multimedia content can be stored in a personal computer. Accordingly, there is an increasing demand for programs that effectively manage multimedia contents in personal computers.

In addition, with the spread of various types of mobile terminals and the provision of various mobile multimedia services, users can easily access multimedia contents anytime and anywhere. Although the storage capacity of mobile terminals is increasing, it is not enough to store and use all the multimedia contents. In addition, it is inconvenient to find and execute desired multimedia contents directly in a mobile terminal having limited I / O resources.

 Despite the rapid development of mobile terminal technology, it is still difficult to access and manage a lot of multimedia contents stored in mobile terminals.

Meanwhile, the personalization system is rapidly emerging as a way to solve the above problems. However, most personalization systems are configured on the server side, and as the number of clients increases, the amount of data to be processed increases.

In addition, the personalization system may have a function of recommending new multimedia content in accordance with each user's registration of his / her preferred genre and taste. However, there is a troublesome problem because the user needs a procedure for registering the function in advance.

It is an object of the present invention to provide a personal multimedia jukebox system and a multimedia recommendation method capable of effectively managing personal multimedia contents.

SUMMARY OF THE INVENTION An object of the present invention is to provide a personal multimedia jukebox system and a multimedia recommendation method capable of easily and effectively recommending multimedia contents suitable for an individual.

An object of the present invention is to provide a personal multimedia jukebox system and a multimedia recommendation method capable of recommending multimedia content suitable for an individual without burden or scalability problems in a server.

An object of the present invention is to provide a personal multimedia jukebox system and a multimedia recommendation method capable of easily and effectively recommending multimedia contents suitable for an individual without a prior registration procedure for an individual's preferred genre and taste.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention, which are not mentioned above, can be understood by the following description, and will be more clearly understood by the embodiments of the present invention. It will also be readily appreciated that the objects and advantages of the invention may be practiced by the means and combinations thereof indicated in the claims.

According to an aspect of the present invention, there is provided a method for recommending multimedia content, wherein the terminal is executed in a terminal having at least a function of searching for multimedia content and storing the multimedia content by accessing a server providing the multimedia content. A first step of acquiring and accumulating weblog information in a process of searching for selecting multimedia contents by accessing the server; A second step of determining, by the terminal, a behavior type of the user, wherein the behavior type is classified into a predetermined number of types by the terminal based on the weblog information accumulated in the first step; And a third step of the terminal recommending the multimedia content based on the behavior type of the user determined in the second step, from among the plurality of multimedia contents recommendable to the user.

According to an aspect of the present invention, there is provided a method for recommending multimedia content, wherein the terminal is executed in a terminal having at least a function of searching for multimedia content and storing the multimedia content by accessing a server providing the multimedia content. Determining a search characteristic of a user based on the weblog information obtained in the process of searching for selecting multimedia content by accessing the server; A second step of the terminal calculating a preference for each multimedia content based on the search characteristic determined in the first step; And a third step of recommending the multimedia content based on the preference calculated in the second step as a main factor among the plurality of multimedia contents that the terminal can recommend to the user.

Agent-based multimedia jukebox system according to an aspect of the present invention, in the agent-based multimedia jukebox system for downloading, storing, and managing the multimedia content from a server providing the multimedia content, the multimedia content downloaded from the server A content storage unit for storing; A collection agent collecting usage information including information generated in the process of browsing or playing the multimedia content; A usage information storage unit for storing the usage information; Among the plurality of multimedia content stored in the content storage unit, characterized in that it comprises a recommendation agent for recommending the multimedia content based on the usage information.

According to an aspect of the present invention, it is possible to effectively manage the multimedia content of the individual.

According to an aspect of the present invention, it is possible to easily and effectively recommend multimedia content suitable for an individual.

According to an aspect of the present invention, it is possible to recommend a multimedia content that is suitable for an individual without burden or scalability on the server.

According to an aspect of the present invention, it is possible to easily and effectively recommend multimedia content suitable for an individual without a prior registration procedure for an individual's preferred genre and taste.

According to an aspect of the present invention, by automatically grasping the behavior type and preference genre of the individual, it is possible to efficiently use the stored content by allowing the individual to download the content highly likely to be preferred to the playback terminal.

1 is a view showing an example of the entire system surrounding the multimedia jukebox system of the present invention.
2 is a diagram illustrating the structure of a website that provides music contents.
3 is a diagram illustrating the internal configuration of the personal multimedia jukebox system 200 in more detail and illustrating the server 100 and the multimedia playback terminal 300 according to an embodiment of the present invention.
4 is a diagram illustrating usage information stored in the usage information storage unit 240.
5 is a flowchart illustrating a method of recommending multimedia content according to an embodiment of the present invention.
6 is a diagram illustrating values (d, v, s) of typical search characteristics in each behavior type.
7 is a flowchart illustrating a method for recommending multimedia content according to another embodiment of the present invention.
8 is a table illustrating situation attributes when playing each multimedia content.
9 is a diagram illustrating a decision tree for genre recommendation according to a situation.

The above objects, features, and advantages will become more apparent from the following detailed description with reference to the accompanying drawings, and as such, those skilled in the art to which the present invention pertains may share the spirit of the present invention. It will be easy to implement. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a view showing an example of the entire system surrounding the multimedia jukebox system of the present invention.

The server 100 has at least a function of providing multimedia content, and may be operated by a content provider and connected through the Internet 400. The user may access the server 100 to access a web page and the like, and may purchase multimedia content existing in the server 100 or search for desired multimedia content.

The hardware configuration of the server 100 may be a general general purpose server configuration, and may store control devices, software, and content such as a CPU capable of responding to the request of the multimedia jukebox system 200 and controlling its own components. A storage space such as an HDD, etc., may be provided with a DBMS system for the efficient management and search of content and information present in the storage space. Of course, you can use only the file system which is another type of management without DBMS system. The multimedia jukebox system 200 which accesses the server 100 searches for a web page made of HTML using a web browser or the like.

The multimedia jukebox system 200 has a function of downloading, storing, managing, playing, and recommending multimedia contents, and at least has a function of accessing the server 100 to search for multimedia contents and to store the multimedia contents. The multimedia jukebox system 200 may be implemented by mounting software on a general-purpose personal computer or by mounting software on a portable terminal. The multimedia jukebox system 200 may be implemented in a fixed or portable manner, and may be implemented in dedicated hardware or non-dedicated hardware. The multimedia jukebox system 200 is connected to the server 100 through a communication network 400 such as the Internet.

The multimedia playback terminal 300 is a terminal that receives and plays multimedia content from the multimedia jukebox system 200.

The multimedia playback terminal 300 may download the multimedia content from the multimedia jukebox system 200 and execute the content. For example, the multimedia playback terminal 300 may be an MP3 player, an audio player, a PMP, a DMB receiver, a PDA, or a navigation terminal.

The communication network 400 is a communication network for transmitting and receiving between the server 100 and the multimedia jukebox system 200, and is a wired communication network such as the Internet, Wi-Fi, Wibro, CDMA, and W-. It may be a wireless communication network such as CDMA or HSPDA.

The short range communication 500 is used for short range transmission and reception between the multimedia jukebox 200 and the multimedia playback terminal 300. Wired communication methods such as USB communication, COM (UART), Parallel Port, Serial ATA (SATA) can be used, and Wi-Fi, Bluetooth, ZigBee Wireless communication schemes such as this may be used.

The multimedia content may be an audio file, a video file, or a combination of these files. In the case of an audio file, the multimedia content may be in a format such as MP3, AAC or PCM. In the case of a video file, such as MPEG4, MPEG2, MPEG1, H.264, VC, etc. May be in a format.

The multimedia jukebox system 200 according to the present invention may download and store multimedia contents from the server 100 and store and manage the multimedia contents. However, the multimedia jukebox system 200 may play the multimedia contents. The multimedia jukebox system 200 of the present invention promotes efficient management of numerous multimedia contents and provides personalized services.

In particular, the multimedia jukebox system 200 of the present invention performs a function of recommending multimedia contents suitable for a user from among a plurality of multimedia contents in order to provide a personalized service to the user.

5 is a flowchart illustrating a method of recommending multimedia content according to an embodiment of the present invention.

The multimedia jukebox system 200 has at least a function of searching for multimedia content by accessing a server 100 providing multimedia content and a function of storing the multimedia content and recommending the same to a user.

The multimedia jukebox system 200 obtains and accumulates weblog information in a process of searching for multimedia content by accessing a webpage provided by the server 100 (S100). The multimedia jukebox system 100 may acquire and collect weblog information using a specialized browser.

Multimedia jukebox system 200 determines the type of behavior of the user based on the accumulated weblog information (S110). In addition, the multimedia jukebox system 200 recommends the multimedia content to the user based on the behavior type as a main factor among the plurality of multimedia contents that can be recommended to the user (S120). The above processes will be described in detail below.

In general, a person acts in a certain way according to his or her unique motivational factors from birth to growth to the present. This tendency is called an action type. The behavior type classifies a user's behavior tendency into a predetermined number of types. There are many ways to classify behavior, but for example, Dr. William Moulton Marston, a professor of psychology at Columbia University, suggests Dominance, D, Influence, I, and Steadiness, S. There are four types, namely "disc type" and "conscientiousness" (type C).

On the other hand, the specialized web browser mounted on the multimedia jukebox system 200 is configured to collect the footsteps weblog from the web page when connecting to the server 100. The web browser is configured to obtain the web log information by using meta tag in Markup Language such as HTML or XHTML. The acquired weblog information is accumulated, and the search characteristics such as search depth, search variety, and staying time of the user can be known from the accumulated weblog information.

2 is a diagram illustrating the structure of a website that provides music contents.

The website structure illustrated in FIG. 2 is composed of three levels: a service page, a service region, and a service element. The service page has a plurality of service regions, and the service region has a plurality of service elements.

The following describes the search depth, the search diversity, and the time of stay, with the service page being 1, the service area being 2, and the service element being 3.

The navigation depth is the number of paths that a user takes to visit a website (main) and purchase level 3 content. For example, the navigation path is main → new releases (level 1) → music A (level 3). ) → genre (level 1) → music B (purchase, level 3), the search depth is 3.

The search diversity refers to the number of times the level has moved to the same level or higher level (level 1> level 2> level 3) on the search path. In the example search path above, music A (level 3) → genre Has a search diversity of 1 in (level 1). The residence time means the total time until the user accesses the website and purchases the content.

In an embodiment of the present invention, the multimedia jukebox system 200 uses the meta tag to obtain information about the search depth, search diversity, and residence time. In order to implement an embodiment of the present invention, a specialized web browser may be used, and the web page provided by the server 100 may include the following code.

You can specify the level in the meta tag of a web page and recognize it in a specialized browser to find out which level you are at. This meta tag is not a information directly expressed to the user as a method of adding information of the corresponding web page, but the meta tag can be used to know the depth of navigation and the diversity of navigation.

A search characteristic vector, wLog, which can be obtained by accumulating log information in a process of searching for a user to purchase or select multimedia content, is defined as the following equation.

The search feature vector is composed of search depth (hereinafter referred to as 'd'), search diversity (hereinafter referred to as 'v') and dwell time (hereinafter referred to as 's'), and this search feature vector, wLog, is a multimedia jukebox Although the system 200 may generate from one search process, it may be generated by accumulating and averaging a plurality of search processes.

On the other hand, the behavior type of the user may be calculated using the distance from the representative vector in each behavior type from the search feature described above.

 6 is a diagram illustrating values (d, v, s) of typical search characteristics in each behavior type.

Pj is a vector representation of the values (d, v, s) of typical search features representing each behavior type.

P _j = (d, v, s), j = {D, I, S, C}

P _D : Representative vector of search characteristics in type D. Each element consists of p _Dk , k = {d, v, s}

P _I : Representative vector of search characteristics of type I, each element consisting of p _Ik , k = {d, v, s}

P _S : Representative vector of search characteristics in S type. Each element consists of p _Sk , k = {d, v, s}

P _C : Representative vector of search characteristics in type C. Each element consists of p _Ck , k = {d, v, s}

When each element of wLog, which is the user's search feature vector, is represented by u _k , k = {d, v, s}, the distance between the user's search feature vector and the representative vector of each behavior type is calculated as follows.

d _D : Distance between user's search characteristic vector (wLog) and representative vector of type _D (P _D )

d _I : Distance between user's search characteristic vector (wLog) and representative vector of type _I (P _I )

d _S : Distance between user's search characteristic vector (wLog) and representative vector of type _S (P _S )

d _C : Distance between user's search characteristic vector (wLog) and representative vector of type _C (P _C )

According to one embodiment of the present invention, the distance between the user's search feature vector (wLog) and a representative vector for each action type may be determined to determine the action type having the shortest distance as the action type of the user. The multimedia content is recommended based on the determined user's behavior type.

{For reference, in another embodiment of the present invention, the preference for each multimedia content is calculated based on the navigation characteristics of the user without explicitly determining the user's behavior type, and the multimedia content is recommended based on the preference as a main factor. There is a method, which will be described later.}

Once the user's behavior type is determined, multimedia content can be recommended in the following ways. Each multimedia content has attributes such as release date, popularity, artist (singer, director) and title.

The release date attribute is important because the D-type is a behavioral type that accepts challenges, makes decisions faster, and is interested in new things. Therefore, the multimedia content is recommended to the user determined as the D type in consideration of the release date attribute or the release date attribute as a main factor.

Type I is a social type, a type of behavior that likes contact with people, respects people's opinions, and follows well, so the popularity of the above four attributes is important. Therefore, the multimedia content is recommended to the user determined as the I type by considering only the popularity attribute or using the popularity attribute as a main factor.

The S type is a stable, consistent, introspective and relaxed type of behavior, so the artist (singer, director) attribute is important among the above four attributes. Therefore, the multimedia content is recommended to the user determined as the S type considering only the number of the same artist (singer, director) or the number of the same artist (singer, director).

The C type is a deliberate and analytic type of behavior, so the number of times a user uses it is important. If the title attributes are the same, the content is determined to be the same multimedia content and the number of times using the same multimedia content is calculated. Accordingly, the multimedia content is recommended considering only the number of times or using the number of times as the main factor.

There may be various ways of associating the determined user's behavior type with the multimedia content to be recommended. In the above, multimedia content is recommended by connecting D type, I type, S type, and C type with release date, popularity, artist (singer, director), and number of times of use, but it may be recommended by connecting other attributes of multimedia content. Of course.

In the above, the method of determining the behavior type of the user and recommending the multimedia content is described. Hereinafter, a method of recommending multimedia content without explicitly determining a user's behavior type as one will be described.

7 is a flowchart illustrating a method for recommending multimedia content according to another embodiment of the present invention.

The multimedia jukebox system 100 obtains and accumulates weblog information in the process of searching for access to the server 100 to select multimedia content (S200). The search characteristic of the user is determined based on the accumulated weblog information (S210).

As described above, the determination of the search characteristics of the user based on the weblog information is the same as in the above-described embodiment, and thus the description thereof is omitted.

The multimedia jukebox system 200 calculates a preference for each multimedia content on the basis of the determined search characteristics of the user (S220). This includes calculating a weight vector from the search characteristics of the user and calculating a preference for each multimedia content by calculating the weight vector and the attribute vector of each multimedia content. This will be described later.

The multimedia jukebox system 200 recommends the multimedia content using the above preference as a main factor (S230). The multimedia jukebox system 200 may recommend multimedia content based only on the preferences, and may consider the other factors as well as the preferences.

The method for obtaining the weight vector from the search characteristics of the user is as follows. The weight vector W is composed of four factors, w _D , w _I , w _S and w _C and represents the weight for each behavior type, that is, the degree to which the user is associated with each behavior type. Each element w _j , j = {D, I, S, C} of the weight vector W can be obtained as follows.

In the above formula, d _j represents the distance between the search feature vector of the user for each behavior type and the representative vector for each behavior type in the above embodiment. That is, it is as follows.

Each multimedia content has its own attributes. Attributes of multimedia content may be defined and used as vectors. The attribute vector AV for the multimedia content has four elements, that is, r, p, m, f, and is defined and used as follows.

r: Relevant to release date, normalized to a value between [0,1]

p is related to popularity and is normalized to a value between [0,1]

m: Total multimedia contents are related to the number of contents of the same artist as the artist of the corresponding multimedia contents and normalized to a value between [0,1].

f: Relevant to the number of times the user uses the content, normalized to a value between [0,1]

As mentioned in the above embodiment, the release date attribute is an important attribute for a D type user, a popularity attribute is an important attribute for an I type user, the number of contents of the same artist is an important attribute for an S type user, and a user's content for a C type user. The number of uses is an important attribute.

Therefore, the preference for each multimedia content is obtained by multiplying the weights associated with each attribute value and adding them together. In other words, the value r is multiplied by the weight w _D , the p value is multiplied by the weight w _I , the m value is multiplied by the weight w _S , and the f value is multiplied by the weight w _C to obtain the sum.

The preference pr may be calculated internally of the weight vector W and the attribute vector AV as in the following equation.

For each multimedia content, the user's preference may be quantified as described above, and the multimedia content may be recommended to the user based on the preference value. When recommending multimedia content, it may be recommended considering only preferences, but may be recommended in connection with other factors.

3 is a diagram illustrating the internal configuration of the personal multimedia jukebox system 200 in more detail and illustrating the server 100 and the multimedia playback terminal 300 according to an embodiment of the present invention.

According to FIG. 3, the multimedia jukebox system includes a collection agent 210, a recommendation agent 220, a transmission agent 230, a usage information storage unit 240, and a content storage unit 250.

4 is a diagram illustrating usage information stored in the usage information storage unit 240.

The weblog information is footstep information that records the user's search in the server 100. The weblog information includes a user ID, a date, a search depth, a search diversity, and a residence time. This information is information acquired in the process of the individual to search for and purchase the multimedia content in the server 100. The weblog information may be information that can be stored for each search. In this case, each weblog information is stored for each search.

The reproduction information is information obtained from the multimedia reproduction terminal 300, and may also include its own reproduction information when the multimedia jukebox system 200 reproduces itself. The reproduction information includes ID (M_id) of the multimedia content, time, date, day of the week, weather, temperature information, and information on the number of times of use of the corresponding content. The attribution information is the attribution information of the multimedia content and includes the multimedia content ID (M_id), genre, artist (singer, director), creator, title, popularity, release date, title, and the like. Weblog information, playback information, and attribute information may be stored using a file system or may be stored using a database system.

The collection agent 210 performs a function of collecting and managing information including the weblog information, the playback information, and the attribute information. The collection agent 210 may collect the information from the multimedia jukebox system 200 itself, or collect the information in cooperation with the server 100 or the multimedia playback terminal 300.

The recommendation agent 220 recommends personalized multimedia content to the user based on the usage information collected by the collection agent 210. The recommendation agent 220 determines the user's behavior type from the weblog information and recommends the multimedia content to the user based on the determined behavior type as a main factor. In another embodiment, the recommendation agent 220 determines the user's search characteristics from the weblog information, calculates the preferences for each multimedia content in consideration of the attributes of the multimedia contents from the search characteristics, and uses the preference as the main factor. Recommend content.

The transmission agent 230 is responsible for the transmission function with the server 100 and the multimedia playback terminal 300. The transmission agent 230 is connected to the server 100 to search for multimedia content, to download the multimedia content from the server 100, to download the multimedia content to the multimedia playback terminal 100, the multimedia playback terminal ( It is involved in the process of receiving playback information from 100).

The collection agent 210, the recommendation agent 220, and the transmission agent 230 may be implemented as dedicated hardware, or may be software components mounted on general-purpose hardware.

In the above description, the recommendation of the multimedia content using the behavior type or the preference as the main factor has been described, but the multimedia content may be recommended in consideration of the other factors together with the above factors. For example, the user may consider the user's preferred genre according to the user's situation.

8 is a table illustrating situation attributes when playing each multimedia content. A situation can be weather, time of day (whether it is dawn, morning, afternoon, evening or night), and temperature.

The reproduction information of FIG. 4 includes contextual information at the time of reproduction for each of the reproduced multimedia contents. The genre of each multimedia content can be known by the attribute information. Then, whenever there is a reproduction of the multimedia content, the situation at that time can be accumulated, and the genre of the multimedia content selected by the user according to the situation can be extracted and built in advance as a decision tree.

9 is a diagram illustrating a decision tree for genre recommendation according to a situation.

In one embodiment of the present invention, information about a user's preferred genre according to a situation is accumulated. The multimedia jukebox system 200 detects the current situation or is provided through a communication network. The time, day of the week, etc. can be known by themselves, and the weather and temperature can be provided through a communication network or can be sensed through its own sensor. After determining the current situation, the genre to be recommended to the user is determined by the decision tree. The recommendation genre is used as an auxiliary factor when recommending multimedia content in the above embodiment, rather than being used alone.

In the above, the multimedia content recommendation method of the present invention has been described based on the multimedia jukebox system 200. However, the multimedia content recommendation method itself is not necessarily the multimedia jukebox system 200, but is connected to the server 100. All terminals that search for multimedia content and store the multimedia content may be performed. For example, if the multimedia reproducing terminal 300 is a terminal capable of directly accessing the server 100 to search for multimedia contents and to store the multimedia contents, the multimedia reproducing terminal 300 may be programmed to execute the above-mentioned multimedia content recommendation method.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. The present invention is not limited to the drawings.

100 Server 200 Multimedia Jukebox System
210 Collecting Agent 220 Recommended Agent
230 Transfer Agent 240 User Information Storage
250 Content storage unit 300 Multimedia playback terminal
400 Network 500 Local Area Communication

Claims (14)

A multimedia content recommendation method executed in a terminal having at least a function of searching for multimedia content and storing multimedia content by accessing a server providing multimedia content,
A first step of acquiring and accumulating weblog information in a process of searching for the multimedia content by the terminal accessing the server;
The terminal extracts a user's search feature vector including search depth, search diversity, and residence time based on the weblog information accumulated in the first step, and extracts the search feature vector and each action type. A second step of determining a behavior type of the user based on the distance between the representative vectors, wherein the behavior type is a classification of the user's behavior tendency into a predetermined number of types;
A third step of, by the terminal, recommending the multimedia content based on the behavior type of the user determined in the second step, from among a plurality of multimedia contents recommendable to the user;
Multimedia content recommendation method comprising a.
A multimedia content recommendation method executed in a terminal having at least a function of searching for multimedia content and storing multimedia content by accessing a server providing multimedia content,
A first step of determining a user's search characteristics including search depth, search diversity, and dwell time, based on the weblog information acquired in the process of the terminal accessing the server to select multimedia contents;
The terminal weights each action type from the search characteristics determined in the first step, wherein the action type is classified into a predetermined number of types of behavior trends of the user, and the weights for each action type are a weight vector. Calculating a preference for each multimedia content by obtaining a dot product between the weight vector and the property vector of each multimedia content, wherein the property vector is a vector representing each property of the multimedia content as a value;
A third step of recommending, by the terminal, the multimedia content based on the preference calculated in the second step as a main factor among a plurality of multimedia contents that can be recommended to the user;
Multimedia content recommendation method comprising a.
delete The method according to claim 1 or 2,
The search depth and the search diversity are extracted based on the level information included in the meta tag of the web page using a specialized browser.
delete delete delete The method according to claim 1 or 2
The action type includes a proactive, social, stable and discreet method.
The method according to claim 1 or 2,
When recommending multimedia content in the third step,
And determining a genre to be recommended based on the contextual information including the current weather.

An agent-based multimedia jukebox system for downloading, storing, and managing multimedia contents from a server providing multimedia contents,
A content storage unit for storing the multimedia content downloaded from the server;
A collection agent collecting usage information including information generated in the process of browsing or playing the multimedia content;
A usage information storage unit for storing the usage information;
And a recommendation agent recommending the multimedia content based on the usage information among the plurality of multimedia contents stored in the content storage unit.
The usage information includes weblog information,
When the recommendation agent recommends multimedia content,
Determine the user's search characteristics, including search depth, search diversity, and dwell time, from the weblog information in the usage information; And classifying the content into a number of types, and recommending multimedia content based on the determined user's behavior type as a main factor.
delete delete delete The method according to claim 10,
Agent-based multimedia jukebox system, characterized in that it further comprises a transport agent comprising at least the function of downloading the multimedia content to the multimedia playback terminal (300).

Images