CN108765076A - Mother and baby's content recommendation method, device and readable storage medium storing program for executing - Google Patents
Mother and baby's content recommendation method, device and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN108765076A CN108765076A CN201810509212.0A CN201810509212A CN108765076A CN 108765076 A CN108765076 A CN 108765076A CN 201810509212 A CN201810509212 A CN 201810509212A CN 108765076 A CN108765076 A CN 108765076A
- Authority
- CN
- China
- Prior art keywords
- user
- infant
- baby
- mother
- attribute
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of mother and baby's content recommendation method of offer of the embodiment of the present invention, device and readable storage medium storing program for executing.This method includes：The insertionization matrix for obtaining user to be predicted indicates, includes the commodity data with the user-association to be predicted in the insertionization matrix expression；Insertionization matrix expression is input in infant's attribute forecast model, exports prediction result, the prediction result includes at least one infant's attribute of the user to be predicted；Recommend corresponding mother and baby's content to the user to be predicted based on the prediction result.Thereby, it is possible to provide matched community content to the user, water conservancy diversion of mother and baby's electric business user to mother and baby community and the interest-degree to community content are promoted, and then improve the clicking rate of community content, cognition of the culture user to contents community.
Description
Technical field
The present invention relates to mother and baby's commending contents field, in particular to a kind of mother and baby's content recommendation method, device and can
Read storage medium.
Background technology
With the development of vertical mother and baby's electric business, the simple purchase function of providing electric business is increasingly limited to the dilatation of category
With the competitiveness of supply chain, and mother and baby's user group it is natural there are highly viscous community attributes, can be with by developing mother and baby community
Promotion user enlivens and APP stay times, preferably back feeding electric business platform.The pattern of content+electric business is increasingly becoming vertical electric business
The standard configuration of industry, the efficiency for improving water conservancy diversion mutually between electric business and contents community are very valuable.
Mother and baby electric business user during doing shopping, cognition and demand to contents community be it is on the weak side, user's
Navigation patterns mainly with electric business merchandising, public praise comment based on, search be also inclined electric business class SKU based on, it is attached to electric business platform
The community content added lacks cognition.How to provide matched community content to the user, promotes mother and baby's electric business user to mother and baby community
Water conservancy diversion, be those skilled in the art's technical problem urgently to be resolved hurrily.
Invention content
In order to overcome above-mentioned deficiency in the prior art, the purpose of the present invention is to provide a kind of mother and baby commending contents sides
Method, device and readable storage medium storing program for executing can provide matched community content to the user, promote mother and baby's electric business user to mother and baby community
Water conservancy diversion and interest-degree to community content, and then improve the clicking rate of community content, cognition of the culture user to contents community.
To achieve the goals above, technical solution used in the embodiment of the present invention is as follows：
In a first aspect, the embodiment of the present invention provides a kind of mother and baby's content recommendation method, it is applied to server, the server
In be configured with infant's attribute forecast model, the method includes：
The insertionization matrix for obtaining user to be predicted indicates, includes and the use to be predicted in the insertionization matrix expression
The associated commodity data in family；
Insertionization matrix expression is input in infant's attribute forecast model, prediction result is exported, it is described
Prediction result includes at least one infant's attribute of the user to be predicted；
Recommend corresponding mother and baby's content to the user to be predicted based on the prediction result.
Optionally, before the insertionization matrix for obtaining user to be predicted the step of, the method further includes：
Configure infant's attribute forecast model, wherein infant's attribute forecast model includes input layer, convolution
Layer, pond layer, full articulamentum and output layer；
The mode of configuration infant's attribute forecast model, including：
Training sample is obtained, the training sample includes the multiple users for having infant's attribute；
The training sample is handled, insertionization matrix to be entered is obtained and indicates；
The insertionization matrix is indicated to be input to the convolutional layer by the input layer；
The insertionization matrix is indicated by the convolutional layer to carry out convolution algorithm, is exported to the pond layer corresponding
Characteristic pattern；
It is output to the full articulamentum after the characteristic pattern is converted to corresponding multi-C vector by the pond layer,
In, the node of the full articulamentum is connect with the node of the pond layer；
The node of the full articulamentum random drop predetermined ratio calculates the characteristic pattern, and result of calculation is defeated
Go out to piecewise linear function and carry out calculation processing, calculation processing result is exported to the output layer；
The calculation processing result is input in softmax classification functions by the output layer calculates each infant category
The probability value of property classification；
In above-mentioned training process, according to the infant of each described other probability value of infant's Attribute class and corresponding user
Attribute calculates cross entropy penalty values, and updates the infant using stochastic gradient descent algorithm according to the cross entropy penalty values
Each layer parameter of attribute forecast model, when the cross entropy penalty values meet training end condition, deconditioning exports mesh
It marks model parameter and calculates and scheme；
Infant's attribute forecast model is obtained according to the objective model parameter and calculating figure.
Optionally, described that the training sample is handled, obtain the step of insertionization matrix to be entered indicates, packet
It includes：
For each user with infant's attribute, the commodity data having with the user-association, the commodity are obtained
Data include goods browse data, articles storage data, the one of which in commodity purchasing data or multiple combinations, the quotient
Product browsing data, articles storage data, commodity purchasing data respectively include trade name and corresponding item property；
The commodity data is filtered, and respectively in filtered commodity data trade name and corresponding quotient
Product attribute carries out numeralization processing, generates dictionary file, and the dictionary file includes trade name coding and the quotient of each commodity
Product attribute coding；
Insertionization matrix to be entered is obtained after handling the dictionary file to indicate.
Optionally, the mode that the commodity data is filtered, including：
The commodity that the occurrence number of trade name and item property in the commodity data is less than to preset times carried out
Filter.
Optionally, it is described the dictionary file is handled after obtain the step of insertionization matrix to be entered indicates,
Including：
The trade name quantity and item property quantity in the dictionary file are counted, statistical result is generated；
The length to be entered of infant's attribute forecast model is generated according to the statistical result；
The insertion of each trade name and item property in the dictionary file is searched from preconfigured word embeded matrix
Vector；
Embedded vector generation based on the length to be entered and each trade name found and item property waits for defeated
The insertionization matrix entered indicates.
Optionally, infant's attribute forecast model includes input layer, convolutional layer, pond layer, full articulamentum and defeated
Go out layer, it is described that insertionization matrix expression is input in infant's attribute forecast model, export the step of prediction result
Suddenly, including：
The insertionization matrix is indicated to be input to the convolutional layer by the input layer；
The insertionization matrix is indicated by the convolutional layer to carry out convolution algorithm, is exported to the pond layer corresponding
Characteristic pattern；
It is output to the full articulamentum after the characteristic pattern is converted to corresponding multi-C vector by the pond layer,
In, the node of the full articulamentum is connect with the node of the pond layer；
The full articulamentum calculates the characteristic pattern using whole nodes, and result of calculation is exported to segmented line
Property function carry out calculation processing after, export corresponding calculation processing result to the output layer；
The calculation processing result is input in softmax classification functions by the output layer calculates each infant category
The probability value of property classification；
The highest infant's attribute of select probability value is as prediction result.
Optionally, described the step of corresponding mother and baby's content is recommended to the user to be predicted based on the prediction result, packet
It includes：
Recommend matched UGC contents and PGC contents to the user based on the prediction result, wherein the PGC contents are closed
It is associated with corresponding infant's attributes section；
The commodity data of each user in corresponding user's set and user's set, and root are obtained based on the prediction result
Commodity data according to each user in corresponding user's set and user's set recommends matched content of good to the user.
Optionally, the commodity data according to each user in the corresponding user set and user's set is to the use
Family is recommended the step of matched content of good, including：
Classified to the commodity data of each user in user set and user set according to confidence threshold,
Obtain classification results；
Recommend matched content of good to the user according to the classification results.
Second aspect, the embodiment of the present invention also provide a kind of mother and baby's content recommendation device, are applied to server, the service
Infant's attribute forecast model is configured in device, described device includes：
Acquisition module, the insertionization matrix for obtaining user to be predicted indicate that the insertionization matrix expression includes
There is the commodity data with the user-association to be predicted；
Input module is exported for insertionization matrix expression to be input in infant's attribute forecast model
Prediction result, the prediction result include at least one infant's attribute of the user to be predicted；
Recommending module, for recommending corresponding mother and baby's content to the user to be predicted based on the prediction result.
The third aspect, the embodiment of the present invention also provide a kind of readable storage medium storing program for executing, are stored in the readable storage medium storing program for executing
Computer program, the computer program, which is performed, realizes above-mentioned mother and baby's content recommendation method.
In terms of existing technologies, the invention has the advantages that：
Mother and baby's content recommendation method, device and readable storage medium storing program for executing provided in an embodiment of the present invention, it is to be predicted by obtaining
The insertionization matrix of user indicates, and insertionization matrix expression is input in infant's attribute forecast model, and output is pre-
It surveys as a result, the prediction result includes at least one infant's attribute of the user to be predicted, then based on the prediction result
Recommend corresponding mother and baby's content to the user to be predicted.Thereby, it is possible to user infant's attributes according to prediction, provide to the user
Matched community content promotes water conservancy diversion of mother and baby's electric business user to mother and baby community and the interest-degree to community content, and then improves
The clicking rate of community content, cognition of the culture user to contents community.
Description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is the application scenarios schematic diagram of mother and baby's content recommendation method provided in an embodiment of the present invention；
Fig. 2 is a kind of flow diagram of mother and baby's content recommendation method provided in an embodiment of the present invention；
Fig. 3 is the structure chart of infant's attribute forecast model provided in an embodiment of the present invention；
Fig. 4 is a kind of structure of the server provided in an embodiment of the present invention for realizing above-mentioned mother and baby's content recommendation method
Block diagram；
Fig. 5 is the functional block diagram of mother and baby's content recommendation device shown in Fig. 4.
Icon：100- servers；110- buses；120- processors；130- storage mediums；140- bus interface；150- nets
Network adapter；160- user interfaces；200- mother and baby's content recommendation device；210- acquisition modules；220- input modules；230- is pushed away
Recommend module；300- user terminals.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Usually herein
The component of the embodiment of the present invention described and illustrated in place's attached drawing can be arranged and be designed with a variety of different configurations.
Therefore, below the detailed description of the embodiment of the present invention to providing in the accompanying drawings be not intended to limit it is claimed
The scope of the present invention, but be merely representative of the present invention selected embodiment.Based on the embodiments of the present invention, this field is common
All other embodiment that technical staff is obtained without creative efforts belongs to the model that the present invention protects
It encloses.
It should be noted that：Similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined, then it further need not be defined and explained in subsequent attached drawing in a attached drawing.Meanwhile the present invention's
In description, term " first ", " second " etc. are only used for distinguishing description, are not understood to indicate or imply relative importance.
Currently, with the development of vertical mother and baby's electric business, the simple purchase function of providing electric business is increasingly limited to category
The competitiveness of dilatation and supply chain, and mother and baby's user group it is natural there are highly viscous community attributes, by developing mother and baby community,
It can promote that user is active and APP stay times, preferably back feeding electric business platform.The pattern of content+electric business is increasingly becoming vertically
The standard configuration of electric business industry, the efficiency for improving water conservancy diversion mutually between electric business and contents community are very valuable.
Inventor has found that a big characteristic of mother and baby community is exactly tenant group, different infants in using community process
The user of attribute has larger purchase and Community demand difference.For example, the user in pregnancy period more pay close attention to maternity dress, exposure suit,
The health care products commodity such as folic acid more pay close attention to the item that the pregnancy period needs to pay attention in mother and baby community, for example, can eat certain foods,
The item that certain variation whether normal, pregnant inspection stages need to pay attention to occurs for body and key index is understood.And 0-1 Sui infant
Mother more pay close attention to breast-feeding, infant's growth indexes, mother new hand must read knowledge etc., milk powder, paper urine are also partial in shopping
Based on the commodity such as trousers.The mother of 1-3 Sui infant more pays close attention to infant cognition culture, can buy and paint sheet, toy, children's garment correlation product
Class, it is more in terms of community to shine babies, tourism partially, based on preschool education class relevant knowledge.The mother of 3-6 Sui infant more pays close attention to above the average age for marriage
The relevant commodity of children's garment, infant nutrient of infant, the more concern big-age-child education of community's aspect, children's activity phase are inside the Pass
Hold.So inventor draws a conclusion, it can preferably help electric business and community to carry out accurate traffic distribution tenant group, carry
Rise user's viscosity.
However, it is found by the inventors that existing in the prior art following problem：Mother and baby's electric business user is in the process done shopping
In, cognition and demand to community be it is on the weak side, the navigation patterns of user mainly based on electric business merchandising, public praise comment,
Search is also to lack to the additional community content of electric business platform and recognize, if being only to provide general society based on inclined electric business class SKU
Area's content does fining matching without being directed to user infant's attribute, most of user to the interest-degree of this some communities content simultaneously
It is not high.In consideration of it, inventor thinks if can be based on the browsing row of user when user browses electric business commodity, public praise
Accurate user portrait, the infant's attribute that the family is portrayed for, collection behavior, Shopping Behaviors, provide the accurately phases such as knowledge, topic
Community content is closed, interest-degree of the electric business user to community content can be improved, and then improves clicking rate of the user to community content,
Cultivate cognition of the user to contents community.
Specifically, the standard configuration of mother and baby community be initial user using when need that infant's attribute information is arranged, it is advantageous in this way
In user is carried out a point group, more accurately content is provided.But on the platform based on electric business, setting infant's attribute is not used
With way, because even providing inlet porting, user fills in that wish is not strong relative to mother and baby's community platform, mother and baby's electric business platform
In default of traffic ingress, it is then less to get infant's attribute information, and obtains less than infant's attribute information for community,
More accurately content can not be just provided and attract user's browsing, form the closed loop of a negative-feedback.To solve the above-mentioned problems, it invents
People has found there is very abundant electric business user behavior on mother and baby's electric business platform under study for action, therefore can be by having infant
The electric business user behavior of attribute is labeled, and not no fill message is estimated out plus new user behavior by deep learning model
User infant's attribute.Meanwhile the community content for having infant's attribute browsing can form resources-type sample mark again, lead to
The infant's attribute for portraying the mapping of this part resource is crossed, and can precisely distribute this partial content to suitable infant's attribute
User, and then promote water conservancy diversion of mother and baby's electric business user to mother and baby community.The present invention proposes that following embodiments are existing to solve as a result,
Many drawbacks present in technology.
Defect present in the above scheme in the prior art, is that inventor is obtaining after putting into practice and carefully studying
As a result, therefore, what the discovery procedure of the above problem and hereinafter the utility model embodiment were proposed regarding to the issue above
Solution all should be the contribution that inventor makes the present invention in process of the present invention.
Referring to Fig. 1, for a kind of application scenarios schematic diagram of mother and baby's content recommendation method provided in an embodiment of the present invention.This
In embodiment, the application scenarios include at least one user terminal 300 and lead to at least one user terminal 300
Believe the server 100 of connection.
In the present embodiment, the user terminal 300 can be PC, smart mobile phone, tablet computer, wearable device
Etc., the present embodiment does not limit this in detail.
According to some embodiments of the present invention, the user terminal 300 may include：Including application processing unit and radio frequency/
The processing unit of digital signals processor；Display screen；It may include secondary or physical bond, the membrane keyboard that is covered on display screen or their group
The keypad of conjunction；SIM card；Can include ROM, RAM, flash memory or their storage arbitrarily combined
Device device；Wi-Fi and/or blue tooth interface；NFC chip connects for radio energy receiving coil, the radio telephone of wireless charging
Mouthful；Electric power management circuit with relevant battery；USB interface and connector；With relevant microphone, loud speaker and earphone jack
Message manage system；And the selectable appurtenances of various cameras, global positioning system, accelerator etc..This
Outside, various user terminal applications can be installed on user terminal 300, user terminal application can be used for allowing using user terminal
300 transmit the order for being suitable for operating with other equipment.This kind of application can be downloaded and installed to user from server 100
It, can also be installed on user terminal 300 in advance in the memory of terminal 300.In embodiments of the present invention, user terminal
Mother and baby's commending contents application is installed, the application of mother and baby's commending contents can give directions user to realize user's registration, Yong Hudeng on 300
The functions such as record, articles storage, commodity purchasing, mother and baby's community content browsing, user's invitation.
In the present embodiment, the server 100 should be understood offer processing, data bank, communications service service point.It lifts
For example, server 100 can refer to the single physical processor with related communication and data storage and document library facility, or
It can refer to the aggregate of networking or the processor gathered, related network and storing unit, and to software and one or more
The application software for the service that document library system and support server 100 are provided is operated.Server 100 can configure
Or it is widely different in performance, but server 100 generally may include one or more central processing unit and storage unit.Clothes
Business device 100 can also be wired or wireless including one or more large-scale storage area equipment, one or more power supplys, one or more
Networking component, one or more input output assemblies or one or more operating systems, such as, Windows Server,
Mac OS X, Unix, Linux, FreeBSD etc..
Further, referring to Fig. 2, a kind of flow for mother and baby's content recommendation method provided in an embodiment of the present invention is illustrated
Figure, the method server 100 shown in Fig. 1 execute.It should be noted that mother and baby's content provided in an embodiment of the present invention
Recommendation method is not limitation with Fig. 2 and particular order as described below, which can be as follows
It realizes：
Step S210, the insertionization matrix for obtaining user to be predicted indicate.
Include the commodity data with the user-association to be predicted, institute in the present embodiment, in the insertionization matrix expression
The user behavior of the user to be predicted can be characterized by stating commodity data, specifically can goods browse data, articles storage data, commodity purchase
Buy the one of which or multiple combinations in data.For example, the sku- of the sku-id of user's electric business browsing, sku attributes id, collection
Id, sku attribute id, purchase sku-id, sku attributes id etc..The sku-id can be understood as specific trade name, institute
It states sku attributes id and can be understood as the item property, such as infant clothes, sku attributes id may include color, ruler
Code, pattern etc..
In detail, before step S210 is further elaborated, first to infant's attribute forecast model
Configuration mode illustrates.Referring to Fig. 3, in the present embodiment, infant's attribute forecast model may include input layer, volume
Lamination, pond layer, full articulamentum and output layer.
First, training sample is obtained, the training sample includes the multiple users for having infant's attribute, the infant
Attribute can be defined and extend according to actual scene, for example, infant's attribute may include：The age of infant, baby
The gender of child, the constellation of infant, the age divide section can be the standby pregnant phase, period of pregnancy, 0-1 Sui, 1-3 Sui, 3-6 Sui, baby
The gender of child is broadly divided into man, female, and the constellation of infant is broadly divided into 12 constellations etc. attribute, does not limit in detail herein
System.
Then, the training sample is handled, obtains insertionization matrix to be entered and indicates.With reference to Fig. 3 pairs
It is illustrated in the generating process that the insertion matrix indicates, it is worth noting that, the insertionization matrix of above-mentioned user to be predicted
The generation of expression can also be realized with reference to following manner.
First, for each user with infant's attribute, the commodity data having with the user-association is obtained.It is described
Commodity data includes goods browse data, articles storage data, the one of which in commodity purchasing data or multiple combinations.Institute
It states goods browse data, articles storage data, commodity purchasing data and respectively includes trade name sku-id and corresponding commodity category
Property sku attributes id.
Then, the commodity data is filtered, and respectively to trade name in filtered commodity data and right
The item property answered carries out numeralization processing, that is, Unified coding is carried out to sku-id and sku attributes id, to sku-id, sku
Attribute id is replaced with its number in dictionary file, is generated sku-id, sku attribute id codings, is quantized, generate word
Allusion quotation file D, therefore the dictionary file includes the trade name coding and item property coding of each commodity.Optionally, to institute
It can be that the occurrence number of trade name and item property in the commodity data is small to state the mode that commodity data is filtered
It is filtered in the commodity of preset times (such as twice), so as to improve the accuracy of training sample.
Then, insertionization matrix to be entered is obtained after handling the dictionary file to indicate.As a kind of implementation
Mode can generate statistical result, and root by counting trade name quantity and item property quantity in the dictionary file
The length to be entered of infant's attribute forecast model is generated according to the statistical result.For example, it is every to count all users
The sku numbers for the duplicate removal that week accesses, sku attribute numbers choose the sku numbers after duplicate removal plus the average length of sku attribute numbers as institute
The length L to be entered of infant's attribute forecast model, can make truncation of the length more than L are stated, length is less than the title of L
With 0 polishing.Then, random initializtion dimension is the word embeded matrix E of D × S, and the dictionary is searched from word embeded matrix E
The embedded vector of each trade name and item property in file D, that is, each sku-id, sku attribute id is passed through it is right
The embedded vector that length in the coding lookup equivalent embeded matrix E answered is S.Finally, based on the length to be entered and looking into
The insertion vector of each trade name and item property that find is combined into the matrix X that dimension is L × S, matrix X namely described
Insertionization matrix to be entered indicates that infant's attribute of the user is as training label.
Then, the insertionization matrix is indicated that X is input to the convolutional layer as input by the input layer, simultaneously
Using infant's attribute of the user as training label.
Then, the insertionization matrix is indicated by the convolutional layer to carry out convolution algorithm, is exported to the pond layer
Corresponding characteristic pattern.As an implementation, by 128 width of the input layer can be 5 by the convolutional layer
The one-dimensional convolution kernel for being 2 with 128 width carries out the convolution algorithm that step-length is 1, wherein the convolution kernel of different in width is each responsible for
The local feature for extracting different range size, when carrying out convolution algorithm, needs indicate to carry out corresponding to the insertionization matrix
Zero padding makes the equal length that the length of the characteristic pattern of output and the insertionization matrix of input indicate.It, can after convolution algorithm
To obtain the characteristic pattern of two L × 128, the characteristic pattern of L × 256 will be formed after two characteristic pattern simple concatenations.
Then, the full connection is output to after the characteristic pattern being converted to corresponding multi-C vector by the pond layer
Layer.The characteristic pattern of L × 256 is converted to the vector of 256 dimensions by maximization pond operation as a result, pond operation reduces spy
The dimension for levying figure, improves the generalization ability of network.Wherein, the node of the full articulamentum and the node of the pond layer connect
It connects.
Then, the node of the full articulamentum random drop predetermined ratio (such as 50%) counts the characteristic pattern
It calculates, and result of calculation is exported to piecewise linear function (ReLU) and carries out calculation processing, calculation processing is exported to the output layer
As a result.Wherein, the calculation formula of the piecewise linear function (ReLU) is f (x)=max (0, x), wherein x is full articulamentum
Output, f (x) are the output of piecewise linear function.
Then, the calculation processing result is input in softmax classification functions and calculates each baby by the output layer
The other probability value of child's Attribute class, wherein the calculation formula of the softmax is
In above-mentioned training process, according to the infant of each described other probability value of infant's Attribute class and corresponding user
Attribute calculates cross entropy penalty values (cross-entropy cost), and is used under stochastic gradient according to the cross entropy penalty values
Drop algorithm updates each layer parameter of infant's attribute forecast model, and item is terminated until the cross entropy penalty values meet training
When part, deconditioning, output objective model parameter and calculating are schemed.As a result, according to the objective model parameter and calculate figure can be with
Obtain infant's attribute forecast model.
After the completion of above-mentioned infant's attribute forecast model configures, indicated in the insertionization matrix for getting user to be predicted
Thing loads the calculating figure and objective model parameter of infant's attribute forecast model first.
Insertionization matrix expression is input in infant's attribute forecast model by step S220, output prediction
As a result.
When implementing, the insertionization matrix is indicated to be input to the convolutional layer by the input layer first, then
The insertionization matrix is indicated by the convolutional layer to carry out convolution algorithm, corresponding characteristic pattern is exported to the pond layer,
And it is output to the full articulamentum after by the pond layer characteristic pattern being converted to corresponding multi-C vector.It is then described
Full articulamentum calculates the characteristic pattern using whole nodes, and result of calculation is exported to piecewise linear function and is counted
After calculation processing, corresponding calculation processing result is exported to the output layer.The output layer inputs the calculation processing result
To calculating each other probability value of infant's Attribute class, and the highest infant of select probability value in softmax classification functions
For attribute as prediction result, the prediction result includes at least one infant's attribute of the user to be predicted.
The concrete operation method of the above process can refer to corresponding in the configuration mode of above-mentioned infant's attribute forecast model
Detailed description, it is no longer repeated herein.
Sku-id, sku attribute id that the user of no infant's attribute is accessed, collected, bought by the present embodiment as a result,
Classification score value is carried out using infant's attribute forecast model to predict to obtain results set P { p1, p2..., pj, by the above results collection
Cooperation is the final classification of user infant's attribute.Such as p1Expression gender attribute is female, p2Indicate that attribute is 0-1 Sui, then using
Family infant's attribute is portrayed precious for 0-1 Sui treasure for woman.
Step S230 recommends corresponding mother and baby's content based on the prediction result to the user to be predicted.
In the present embodiment, mother and baby's content can be divided into two kinds of forms：Class content and consumer content are produced, class content is produced
Namely UGC contents and PGC contents, the content that consumer content namely user actively browse, thumb up, collecting.As a kind of implementation
Mode can be based on the prediction result and recommend matched UGC contents and PGC contents to the user after obtaining prediction result.
UGC (User Generated Content) content refer to user's original content namely other with infant
The user of attribute is when mother and baby community issues corresponding UGC contents, if infant's attribute of the user and above-mentioned use to be predicted
Infant's attribute at family is similar, and the UGC content push which issues can be given the use of the user to be predicted by server 100
Family terminal 300.
PGC (Professional Generated Content) content refers to professional production content.The PGC contents
It is associated with corresponding infant's attributes section, is belonged to that is, associated specialist can choose matched infant when creating PGC contents
Property, if the infant's attribute chosen is similar to above-mentioned infant's attribute of user to be predicted, server 100 can will be described
PGC content push gives the user terminal 300 of the user to be predicted.Thus, it is possible to push accurately UGC for the user to be predicted
Content and PGC contents.
And for consumer content, the prediction result can be based on and obtained in corresponding user's set and user's set
The commodity data of each user, and according to the commodity data of each user in corresponding user's set and user's set to this
User recommends matched content of good.For example, all user's set U { u of post-consumer content can be obtained1, u2…unBaby children
Youngster attribute P { pu1, pu2…pun, p is obtained according to confidence threshold μuiCredible classification results set C { c │ c>μ }, then basis is divided
Class result recommends matched content of good to the user.
As a result, through the above steps, the present embodiment can provide matching to the user according to user infant's attribute of prediction
Community content, promote water conservancy diversion of mother and baby's electric business user to mother and baby community and the interest-degree to community content, and then improve community
The clicking rate of content, cognition of the culture user to contents community.
Referring to Fig. 4, a kind of structural schematic block diagram of the server 100 provided for present pre-ferred embodiments.Such as Fig. 4 institutes
Show, the server 100 can make general bus architecture to realize by bus 110.According to the specific of server 100
Using with overall design constraints condition, bus 110 may include any number of interconnection bus and bridge joint.Bus 110 will be various
It is electrically connected to together, these circuits include processor 120, storage medium 130 and bus interface 140.Optionally, server
100 can use bus interface 140 to wait network adapter 150 connects via bus 110.Network adapter 150 can be used for reality
The signal processing function of physical layer in existing cordless communication network, and sending and receiving for radiofrequency signal is realized by antenna.User
Interface 160 can connect external equipment, such as：Keyboard, display, mouse or control stick etc..Bus 110 can also connect respectively
The other circuits of kind, such as timing source, peripheral equipment, voltage regulator or management circuit, these circuits are this field institutes
It is well known, therefore be no longer described in detail.
It can replace, server 100 may also be configured to generic processing system, such as be commonly referred to as chip, the general procedure
System includes：The one or more microprocessors of processing function are provided, and at least part of outer of storage medium 130 is provided
Portion's memory, it is all these all to be linked together by external bus architecture and other support circuits.
Alternatively, server 100 can be realized using following：With processor 120, bus interface 140, Yong Hujie
The ASIC (application-specific integrated circuit) of mouth 160；And it is integrated at least part of the storage medium 130 in one single chip, alternatively,
Server 100 can be realized using following：One or more FPGA (field programmable gate array), PLD (programmable logic devices
Part), controller, state machine, gate logic, discrete hardware components, any other suitable circuit or to be able to carry out the present invention logical
The arbitrary combination of the circuit of various functions described in.
Wherein, processor 120 is responsible for bus 110 and general processing and (including executes and be stored on storage medium 130
Software).Processor 120 can be realized using one or more general processors and/or application specific processor.Processor 120
Example includes microprocessor, microcontroller, dsp processor and the other circuits for being able to carry out software.It should be by software broadly
Be construed to indicate instruction, data or its it is arbitrary combine, regardless of being called it as software, firmware, middleware, microcode, hard
Part description language or other.
Storage medium 130 is illustrated as detaching with processor 120 in Fig. 4, however, those skilled in the art be easy to it is bright
In vain, storage medium 130 or its arbitrary portion can be located at except server 100.For example, storage medium 130 may include passing
Defeated line, the carrier waveform modulated with data, and/or the computer product that is separated with radio node, these media can be by
Processor 120 is accessed by bus interface 140.Alternatively, storage medium 130 or its arbitrary portion are desirably integrated into processing
In device 120, for example, it may be cache and/or general register.
The processor 120 can perform above-described embodiment, can specifically, in the storage medium 130 be stored with described
Mother and baby's content recommendation device 200, the processor 120 can be used for executing mother and baby's content recommendation device 200.
Further, referring to Fig. 5, mother and baby's content recommendation device 200 may include：
Acquisition module 210, the insertionization matrix for obtaining user to be predicted indicate that the insertionization matrix wraps in indicating
Include the commodity data with the user-association to be predicted；
Input module 220, it is defeated for insertionization matrix expression to be input in infant's attribute forecast model
Go out prediction result, the prediction result includes at least one infant's attribute of the user to be predicted；
Recommending module 230, for recommending corresponding mother and baby's content to the user to be predicted based on the prediction result.
It is understood that the concrete operation method of each function module in the present embodiment can refer to above method embodiment
The detailed description of middle corresponding steps, it is no longer repeated herein.
The embodiment of the present invention also provides a kind of readable storage medium storing program for executing, and computer journey is stored in the readable storage medium storing program for executing
Sequence, the computer program, which is performed, realizes above-mentioned mother and baby's content recommendation method.
In conclusion mother and baby's content recommendation method, device and readable storage medium storing program for executing provided in an embodiment of the present invention, by obtaining
It takes the insertionization matrix of user to be predicted to indicate, and insertionization matrix expression is input to infant's attribute forecast model
In, prediction result is exported, the prediction result includes at least one infant's attribute of the user to be predicted, is then based on described
Prediction result recommends corresponding mother and baby's content to the user to be predicted.Thereby, it is possible to user infant's attributes according to prediction, are
User provides matched community content, promotes water conservancy diversion of mother and baby's electric business user to mother and baby community and the interest-degree to community content,
And then the clicking rate of community content is improved, cognition of the culture user to contents community.
In embodiment provided by the present invention, it should be understood that disclosed device and method, it can also be by other
Mode realize.Device and method embodiment described above is only schematical, for example, the flow chart in attached drawing and frame
Figure shows the system frame in the cards of the system of multiple embodiments according to the present invention, method and computer program product
Structure, function and operation.In this regard, each box in flowchart or block diagram can represent a module, section or code
A part, the part of the module, section or code includes one or more for implementing the specified logical function
Executable instruction.It should also be noted that at some as in the realization method replaced, the function of being marked in box can also be with not
It is same as the sequence marked in attached drawing generation.For example, two continuous boxes can essentially be basically executed in parallel, they have
When can also execute in the opposite order, this is depended on the functions involved.It is also noted that in block diagram and or flow chart
Each box and the box in block diagram and or flow chart combination, the special of function or action as defined in executing can be used
Hardware based system realize, or can realize using a combination of dedicated hardware and computer instructions.
In addition, each function module in each embodiment of the present invention can integrate to form an independent portion
Point, can also be modules individualism, can also two or more modules be integrated to form an independent part.
It can replace, can be realized wholly or partly by software, hardware, firmware or its arbitrary combination.When
When using software realization, can entirely or partly it realize in the form of a computer program product.The computer program product
Including one or more computer instructions.It is all or part of when loading on computers and executing the computer program instructions
Ground is generated according to the flow or function described in the embodiment of the present invention.The computer can be all-purpose computer, special purpose computer,
Computer network or other programmable devices.The computer instruction can store in a computer-readable storage medium, or
Person is transmitted from a computer readable storage medium to another computer readable storage medium, for example, the computer instruction
Wired (such as coaxial cable, optical fiber, digital subscriber can be passed through from a web-site, computer, server or data center
Line (DSL)) or wireless (such as infrared, wireless, microwave etc.) mode to another web-site, computer, server or data
Center is transmitted.The computer readable storage medium can be that any usable medium that computer can access either is wrapped
The data storage devices such as server, the data center integrated containing one or more usable mediums.The usable medium can be magnetic
Property medium, (for example, floppy disk, hard disk, tape), optical medium (for example, DVD) or semiconductor medium (such as solid state disk
Solid State Disk (SSD)) etc.
It should be noted that herein, term " including ", " including " or its any other variant are intended to non-row
Its property includes, so that the process, method, article or equipment including a series of elements includes not only those elements, and
And further include the other elements being not explicitly listed, or further include for this process, method, article or equipment institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence " including one ... ", it is not excluded that including institute
State in the process, method, article or equipment of element that there is also other identical elements.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Profit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent requirements of the claims
Variation is included within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.
Claims (10)
1. a kind of mother and baby's content recommendation method, which is characterized in that be applied to server, belong to configured with infant in the server
Property prediction model, the method includes：
The insertionization matrix for obtaining user to be predicted indicates, includes in the insertionization matrix expression and the user to be predicted is closed
The commodity data of connection；
Insertionization matrix expression is input in infant's attribute forecast model, prediction result, the prediction are exported
As a result include at least one infant's attribute of the user to be predicted；
Recommend corresponding mother and baby's content to the user to be predicted based on the prediction result.
2. mother and baby's content recommendation method according to claim 1, which is characterized in that obtain the embedding of user to be predicted described
Before the step of entering matrix, the method further includes：
Configure infant's attribute forecast model, wherein infant's attribute forecast model include input layer, convolutional layer,
Pond layer, full articulamentum and output layer；
The mode of configuration infant's attribute forecast model, including：
Training sample is obtained, the training sample includes the multiple users for having infant's attribute；
The training sample is handled, insertionization matrix to be entered is obtained and indicates；
The insertionization matrix is indicated to be input to the convolutional layer by the input layer；
The insertionization matrix is indicated by the convolutional layer to carry out convolution algorithm, corresponding feature is exported to the pond layer
Figure；
It is output to the full articulamentum after the characteristic pattern is converted to corresponding multi-C vector by the pond layer, wherein
The node of the full articulamentum is connect with the node of the pond layer；
The node of the full articulamentum random drop predetermined ratio calculates the characteristic pattern, and by result of calculation export to
Piecewise linear function carries out calculation processing, and calculation processing result is exported to the output layer；
The calculation processing result is input in softmax classification functions and calculates each infant's Attribute class by the output layer
Other probability value；
In above-mentioned training process, according to infant's attribute of each described other probability value of infant's Attribute class and corresponding user
Cross entropy penalty values are calculated, and infant's attribute is updated using stochastic gradient descent algorithm according to the cross entropy penalty values
Each layer parameter of prediction model, when the cross entropy penalty values meet training end condition, deconditioning exports target mould
Shape parameter and calculating are schemed；
Infant's attribute forecast model is obtained according to the objective model parameter and calculating figure.
3. mother and baby's content recommendation method according to claim 2, which is characterized in that it is described to the training sample at
Reason obtains the step of insertionization matrix to be entered indicates, including：
For each user with infant's attribute, the commodity data having with the user-association, the commodity data are obtained
Including the one of which or multiple combinations in goods browse data, articles storage data, commodity purchasing data, the commodity are clear
Looking at data, articles storage data, commodity purchasing data respectively includes trade name and corresponding item property；
The commodity data is filtered, and respectively in filtered commodity data trade name and corresponding commodity category
Property carry out numeralization processing, generate dictionary file, the dictionary file includes the trade name coding and commodity category of each commodity
Property coding；
Insertionization matrix to be entered is obtained after handling the dictionary file to indicate.
4. mother and baby's content recommendation method according to claim 3, which is characterized in that described to be carried out to the commodity data
The mode of filter, including：
The commodity that the occurrence number of trade name and item property in the commodity data is less than to preset times are filtered.
5. mother and baby's content recommendation method according to claim 3, which is characterized in that it is described to the dictionary file at
The step of insertionization matrix to be entered indicates is obtained after reason, including：
The trade name quantity and item property quantity in the dictionary file are counted, statistical result is generated；
The length to be entered of infant's attribute forecast model is generated according to the statistical result；
The embedded vector of each trade name and item property in the dictionary file is searched from preconfigured word embeded matrix；
Embedded vector based on the length to be entered and each trade name found and item property generates to be entered
Insertionization matrix indicates.
6. mother and baby's content recommendation method according to claim 1, which is characterized in that infant's attribute forecast model packet
Include input layer, convolutional layer, pond layer, full articulamentum and output layer, it is described insertionization matrix expression is input to it is described
In infant's attribute forecast model, export prediction result the step of, including：
The insertionization matrix is indicated to be input to the convolutional layer by the input layer；
The insertionization matrix is indicated by the convolutional layer to carry out convolution algorithm, corresponding feature is exported to the pond layer
Figure；
It is output to the full articulamentum after the characteristic pattern is converted to corresponding multi-C vector by the pond layer, wherein
The node of the full articulamentum is connect with the node of the pond layer；
The full articulamentum calculates the characteristic pattern using whole nodes, and result of calculation is exported to piecewise linear function
After number carries out calculation processing, corresponding calculation processing result is exported to the output layer；
The calculation processing result is input in softmax classification functions and calculates each infant's Attribute class by the output layer
Other probability value；
The highest infant's attribute of select probability value is as prediction result.
7. mother and baby's content recommendation method according to claim 1, which is characterized in that it is described based on the prediction result to this
User to be predicted recommends the step of corresponding mother and baby's content, including：
Recommend matched UGC contents and PGC contents to the user based on the prediction result, wherein the PGC contents are associated with
Corresponding infant's attributes section；
The commodity data of each user in corresponding user's set and user's set is obtained based on the prediction result, and according to institute
The commodity data for stating each user in corresponding user's set and user's set recommends matched content of good to the user.
8. mother and baby's content recommendation method according to claim 7, which is characterized in that described to be collected according to the corresponding user
The step of recommending matched content of good to the user with the commodity data of each user in user's set is closed, including：
Classified to the commodity data of each user in user set and user set according to confidence threshold, is obtained
Classification results；
Recommend matched content of good to the user according to the classification results.
9. a kind of mother and baby's content recommendation device, which is characterized in that be applied to server, belong to configured with infant in the server
Property prediction model, described device include：
Acquisition module, the insertionization matrix for obtaining user to be predicted indicate, include in insertionization matrix expression and
The commodity data of the user-association to be predicted；
Input module, for insertionization matrix expression to be input in infant's attribute forecast model, output prediction
As a result, the prediction result includes at least one infant's attribute of the user to be predicted；
Recommending module, for recommending corresponding mother and baby's content to the user to be predicted based on the prediction result.
10. a kind of readable storage medium storing program for executing, which is characterized in that be stored with computer program, the meter in the readable storage medium storing program for executing
Calculation machine program is performed the mother and baby's content recommendation method realized described in any one of claim 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810509212.0A CN108765076B (en) | 2018-05-24 | 2018-05-24 | Mother and infant content recommendation method and device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810509212.0A CN108765076B (en) | 2018-05-24 | 2018-05-24 | Mother and infant content recommendation method and device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108765076A true CN108765076A (en) | 2018-11-06 |
| CN108765076B CN108765076B (en) | 2021-08-03 |
Family
ID=64005714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810509212.0A Active CN108765076B (en) | 2018-05-24 | 2018-05-24 | Mother and infant content recommendation method and device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108765076B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119935A (en) * | 2019-03-20 | 2019-08-13 | 中国空气动力研究与发展中心高速空气动力研究所 | A kind of science and technology item declares method for procedure tracking and device |
| CN111274497A (en) * | 2020-01-22 | 2020-06-12 | 北京百度网讯科技有限公司 | Community recommendation and model training method and device, electronic equipment and storage medium |
| CN111292168A (en) * | 2020-02-06 | 2020-06-16 | 腾讯科技（深圳）有限公司 | Data processing method, device and equipment |
| CN111861178A (en) * | 2020-07-13 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Service matching model training method, service matching method, device and medium |
| CN112506975A (en) * | 2020-11-30 | 2021-03-16 | 橙脑教育科技(上海)有限公司 | Information recommendation method, device and system |
| CN115619046A (en) * | 2022-12-05 | 2023-01-17 | 湖南工商大学 | Commercial configuration method for community life circle |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105868317A (en) * | 2016-03-25 | 2016-08-17 | 华中师范大学 | Digital education resource recommendation method and system |
| CN106874522A (en) * | 2017-03-29 | 2017-06-20 | 珠海习悦信息技术有限公司 | Information recommendation method, device, storage medium and processor |
| CN107330115A (en) * | 2017-07-12 | 2017-11-07 | 广东工业大学 | A kind of information recommendation method and device |
| CN107766439A (en) * | 2017-09-21 | 2018-03-06 | 汉鼎宇佑互联网股份有限公司 | A kind of personalized recommendation method of fusion structure feature and implicit feedback |
| CN107992531A (en) * | 2017-11-21 | 2018-05-04 | 吉浦斯信息咨询（深圳）有限公司 | News personalization intelligent recommendation method and system based on deep learning |
-
2018
- 2018-05-24 CN CN201810509212.0A patent/CN108765076B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105868317A (en) * | 2016-03-25 | 2016-08-17 | 华中师范大学 | Digital education resource recommendation method and system |
| CN106874522A (en) * | 2017-03-29 | 2017-06-20 | 珠海习悦信息技术有限公司 | Information recommendation method, device, storage medium and processor |
| CN107330115A (en) * | 2017-07-12 | 2017-11-07 | 广东工业大学 | A kind of information recommendation method and device |
| CN107766439A (en) * | 2017-09-21 | 2018-03-06 | 汉鼎宇佑互联网股份有限公司 | A kind of personalized recommendation method of fusion structure feature and implicit feedback |
| CN107992531A (en) * | 2017-11-21 | 2018-05-04 | 吉浦斯信息咨询（深圳）有限公司 | News personalization intelligent recommendation method and system based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| FUZHENG ZHANG等: "Collaborative Knowledge Base Embedding for Recommender Systems", 《KDD ’16》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110119935A (en) * | 2019-03-20 | 2019-08-13 | 中国空气动力研究与发展中心高速空气动力研究所 | A kind of science and technology item declares method for procedure tracking and device |
| CN111274497A (en) * | 2020-01-22 | 2020-06-12 | 北京百度网讯科技有限公司 | Community recommendation and model training method and device, electronic equipment and storage medium |
| CN111274497B (en) * | 2020-01-22 | 2023-08-18 | 北京百度网讯科技有限公司 | Community recommendation and model training method and device, electronic equipment and storage medium |
| CN111292168A (en) * | 2020-02-06 | 2020-06-16 | 腾讯科技（深圳）有限公司 | Data processing method, device and equipment |
| CN111861178A (en) * | 2020-07-13 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Service matching model training method, service matching method, device and medium |
| CN111861178B (en) * | 2020-07-13 | 2024-06-07 | 北京嘀嘀无限科技发展有限公司 | Training method of service matching model, service matching method, equipment and medium |
| CN112506975A (en) * | 2020-11-30 | 2021-03-16 | 橙脑教育科技(上海)有限公司 | Information recommendation method, device and system |
| CN115619046A (en) * | 2022-12-05 | 2023-01-17 | 湖南工商大学 | Commercial configuration method for community life circle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108765076B (en) | 2021-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108765076A (en) | Mother and baby's content recommendation method, device and readable storage medium storing program for executing | |
| CN109166017A (en) | Method for pushing, device, computer equipment and storage medium based on reunion class | |
| CN110209807A (en) | A kind of method of event recognition, the method for model training, equipment and storage medium | |
| CN107480187A (en) | User's value category method and apparatus based on cluster analysis | |
| CN108352035A (en) | Dynamic topology system and method for efficient claims treatment | |
| CN106022856A (en) | Data display method and device | |
| CN107301592A (en) | The method and device excavated for commodity substitute | |
| CN108122128A (en) | A kind of discount coupon method for release management and device | |
| CN107145485A (en) | Method and apparatus for compressing topic model | |
| CN112528616B (en) | Service form generation method and device, electronic equipment and computer storage medium | |
| CN109783749A (en) | A kind of Material for design intelligent recommendation method, apparatus and terminal device | |
| CN103337027A (en) | Clothing information pushing method and system | |
| CN118132752B (en) | Commodity description word classification method and device | |
| CN107832338A (en) | A kind of method and system for identifying core product word | |
| CN109978650A (en) | Construct the method and system of decision tree | |
| Razzaq et al. | An automatic determining food security status: machine learning based analysis of household survey data | |
| CN110852785A (en) | User grading method, device and computer readable storage medium | |
| CN108984688A (en) | Mother and baby's knowledge topic recommended method and device | |
| CN107392259A (en) | The method and apparatus for building unbalanced sample classification model | |
| CN108062385A (en) | The method and system of Users' Interests Mining | |
| CN112380299A (en) | Relational network construction method, device and storage medium | |
| CN107274247A (en) | Wisdom based on cloud computing perceives recommendation method | |
| Xu et al. | Resampling-based noise correction for crowdsourcing | |
| CN112633988A (en) | User product recommendation method and device, electronic equipment and readable storage medium | |
| CN104281599A (en) | Method and device for recommending information to user in social network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220920 Address after: 311100 rooms 809 and 810, building 8, Xixi bafangcheng, Wuchang Street, Yuhang District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou Mitu network technology (Group) Co.,Ltd. Address before: 100000 room 1606, 16th floor, building 1, yard 16, Taiyanggong Middle Road, Chaoyang District, Beijing Patentee before: MIYA BAOBEI (BEIJING) NETWORK TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right |