CN107807935B - Using recommended method and device - Google Patents
Using recommended method and device Download PDFInfo
- Publication number
- CN107807935B CN107807935B CN201610814453.7A CN201610814453A CN107807935B CN 107807935 B CN107807935 B CN 107807935B CN 201610814453 A CN201610814453 A CN 201610814453A CN 107807935 B CN107807935 B CN 107807935B
- Authority
- CN
- China
- Prior art keywords
- application
- user
- group
- reference feature
- correlation
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of application recommended method and devices, belong to network technique field.The described method includes: obtaining the correlation of each application and the reference feature of multiple users in multiple applications；The multiple application is grouped according to the correlation, obtains first using group, the first application group includes the application that multiple correlations are greater than preset threshold；Group is applied for first, the reference feature based on the multiple user establishes corresponding first prediction model, and first prediction model is used to determine the recommendation probability of application based on reference feature；Reference feature and first prediction model based on user to be recommended recommend application to the user to be recommended.The present invention can be improved to the accuracy rate in the first application group using recommendation, so that the application recommended more meets the downloading demand of user, so as to improve the conversion ratio of application.
Description
Technical field
The present invention relates to network technique field, in particular to a kind of application recommended method and device.
Background technique
With the continuous development of network technology, the service that service provider is provided a user by network is also more and more, and more
Come more perfect.For example, when user wants downloading in application, can apply download platform or application by input search key
The corresponding application of service platform downloading.
It, can also be to user using download platform or application service platform other than it can provide the download function of application
Recommend application, specific recommended method can be with are as follows: analyzes user characteristics, pointedly to apply to different users
Recommend, which includes age, gender, place city, interactive information and history Download History etc. on line.For example, for
Finance and money management class application, can be according to the user characteristics such as age of user and place city, it is determined whether recommended, when any use
The age at family at 24 years old or more, and the user where city be city above county level when, recommend the finance and money management class to the user
Using.
In the implementation of the present invention, the inventor finds that the existing technology has at least the following problems:
User characteristics by being analyzed will lead to there are limitation and apply the accuracy rate of recommendation low, recommended
Using and do not meet user intention, user will not download the application of recommendation, so as to cause recommend application conversion ratio it is low.
Summary of the invention
In order to solve problems in the prior art, the embodiment of the invention provides a kind of application recommended method and devices.It is described
Technical solution is as follows:
On the one hand, it provides a kind of using recommended method, which comprises
Obtain the correlation of each application and the reference feature of multiple users in multiple applications；
The multiple application is grouped according to the correlation, obtains first using group, the first application group packet
Include the application that multiple correlations are greater than first threshold；
Group is applied for first, the reference feature based on the multiple user establishes corresponding first prediction model, described
First prediction model is used to determine the recommendation probability of application based on reference feature；
Reference feature and first prediction model based on user to be recommended recommend application to the user to be recommended.
On the other hand, provide a kind of using recommendation apparatus, described device includes:
Correlation obtains module, related to the reference feature of multiple users for obtaining each application in multiple applications
Property；
Grouping module, for according to the correlation obtain the correlation that gets of module to it is the multiple apply into
Row grouping obtains first using group, and the first application group includes the application that multiple correlations are greater than preset threshold；
Model building module, first for obtaining for the grouping module applies group, based on the multiple user's
Reference feature establishes corresponding first prediction model, and first prediction model is used to determine the recommendation of application based on reference feature
Probability；
Recommending module, described first for reference feature and model building module foundation based on user to be recommended
Prediction model recommends application to the user to be recommended.
Technical solution provided in an embodiment of the present invention has the benefit that
By obtaining for indicating using the correlation with the degree of correlation of the reference feature of user, and according to multiple applications
In it is each application and the reference feature of multiple users correlation, to the reference feature degree of correlation of user it is high first application
Group, the first prediction model established using the reference feature based on multiple users obtain each application in the first application group
Recommendation probability carries out can be improved the accuracy rate in the first application group using recommendation using recommendation according to the recommendation probability,
So that the application recommended more meets the downloading demand of user, so as to improve the conversion ratio of application.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is a kind of application environment schematic diagram using recommended method provided in an embodiment of the present invention；
Fig. 2A is a kind of application recommended method flow chart provided in an embodiment of the present invention；
Fig. 2 B is a kind of prediction model Establishing process figure provided in an embodiment of the present invention；
Fig. 3 is a kind of application recommendation apparatus block diagram provided in an embodiment of the present invention；
Fig. 4 is a kind of structural schematic diagram of device 400 provided in an embodiment of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Formula is described in further detail.
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistented with the present invention.On the contrary, they be only with it is such as appended
The example of device and method being described in detail in claims, some aspects of the invention are consistent.
Fig. 1 is a kind of application environment schematic diagram using recommended method provided in an embodiment of the present invention, as described in Figure 1,
The structural schematic diagram of application environment involved in the embodiment of the present invention is shown, this includes: using the application environment of recommended method
Server 101 and at least one terminal 102.
Terminal 102 is connected by wireless or cable network and server 101, and terminal 102 can be computer, intelligent hand
The electronic equipments such as machine, tablet computer.
Server 101 can be the Internet application server, which can mention for Internet application
For background service.Internet application provides the information exchanges service such as voice, video, picture, text as one for intelligent terminal
Application program, have many advantages, such as that voice, video, picture and text can be sent across common carrier, spanning operation system platform.
The Internet application server can be configured as one and provide the server of service by internet, which answers
It can be social application server with server, for example, the corresponding clothes of the social network sites such as instant communication server, forum or microblogging
Business device, can also be the server that can be realized the business such as payment by internet, and the embodiment of the present invention takes Internet application
The type of business device is without specifically limiting.
Certainly, the server 101 or other servers, such as multimedia resource shared server, the present invention are real
Example is applied to be not especially limited the type of the server.
Fig. 2A is a kind of application recommended method flow chart provided in an embodiment of the present invention, and the executing subject of this method is service
Device, A referring to fig. 2, this method comprises:
201, the correlation of each application and the reference feature of multiple users in multiple applications is obtained.
The reference feature is used to indicate that the creditworthiness of user, the reference feature can to use reference fraction representation, user's
Reference score can be determined according to behavioral data on the line of user, can also be determined according to behavioral data is associated under the line of user,
Or determined according to behavioral data is associated under behavioral data on the line of user and line, it can also be determined according to other data of user
The credit score, the present invention is not especially limit this.
It should be noted that behavioral data may include social interaction behavioral data, virtual increment clothes on the line of the user
At least one of the data, economic behaviour data and amusement and leisure behavioral data of being engaged in data, association behavioral data can wrap under line
Include wearable device data, tourism trip data, in O2O (Online to Offline, i.e., under line on line) service for life data
At least one data, under behavioral data and line be associated with behavioral data respectively certainly, on the line and can also include or can be it
His data, the present invention is not especially limit this.
The correlation is used to indicate to apply the degree of correlation with the reference feature of user, and the numerical value of the correlation is bigger, table
Show that application is bigger with the degree of correlation of the reference feature of user, numerical value is smaller, indicates that application is related to the reference feature of user
Degree is smaller.It is more demanding to the creditworthiness of user for example, for financial class application, correspondingly, the finance class apply with
The degree of correlation of the reference feature at family is higher, and accessed correlation is bigger.
In embodiments of the present invention, each application and the correlation of the reference feature of multiple users in multiple applications are obtained
Method includes step 2011 to 2013:
2011, for any application in multiple application, according to the reference feature of multiple user, to multiple user
It is grouped.
According to the reference feature of multiple user or the distribution situation of the reference feature of multiple user, to multiple use
Family is grouped, and the number of users of every group of user can be made identical, can also make the number of users of every two groups of users in multiple groups user
Within a preset range, which can be determined as any fixed value to difference between amount, and the embodiment of the present invention does not make this
It is specific to limit.
By taking the number of users of every group of user is identical as an example, specific group technology can be with are as follows: if should with reference fraction representation
Multiple user is arranged successively by the sequence of reference score from low to high or from high to low, is with preset quantity by reference feature
Benchmark is grouped the multiple users being arranged in order.
The preset quantity can be determined as any fixed value, can also be according to the number of users determination of multiple user and group
Number determines, i.e., the ratio of the number of users of multiple user and group number is determined as the preset quantity, it is, of course, also possible to according to it
His method determines the preset quantity, and the present invention is not especially limit this.
2012, according to every group of user to the downloading situation of the application and multiple user to the downloading situation of the application, obtain
Take the correlation of the application with the reference feature of every group of user.
It can be the number of users for downloading the application and the number of users for not downloading the application to the downloading situation of the application
Ratio；Or the downloading total degree to the application, or to download the number of users of the application and not downloading the application
Other downloading situations such as number of users, the embodiment of the present invention is not construed as limiting the specific meaning of the downloading situation.Correspondingly, root
According to the difference of meaning indicated by the downloading situation, the method for obtaining the application with the correlation of the reference feature of every group of user
Difference, the embodiment of the present invention are also not construed as limiting specific acquisition methods.
According to the downloading situation of application, the correlation of application with the reference feature of user is obtained, capable of embodying downloading, this is answered
The creditworthiness of user, and then can embody using the requirement height to user credit degree.For example, being answered for financial class
With indicating the credit of the application with user when the number of users that the higher grouping user of creditworthiness downloads the application is more
The correlation of degree is higher, i.e., higher with the correlation of the reference feature of user；And for amusement and leisure class application, if every group
The number of users that user downloads the application is not much different, then it represents that the application is lower with the correlation of the reference degree of user, i.e.,
It is lower with the correlation of the reference feature of user.
When the downloading situation of the application is to download the number of users of the application and do not download the number of users of the application, obtain
Take the method for the application and the correlation of the reference feature of every group of user can be with are as follows:
The correlation of the application with the reference feature of every group of user is obtained according to the following formula,
Wherein, i=1,2 ..., n, n indicate the user group number obtained after being grouped to multiple user, IViIndicating should
Using the correlation of the reference feature with i-th group of user, GiIndicate the number of users that the application is downloaded in i-th group of user, BiTable
Show the number of users for not downloading the application in i-th group of user, GTIndicate the number of users that the application is downloaded in multiple user,
BTIndicate that the number of users for not downloading the application in multiple user, " ln " indicate that logarithm operation symbol, " * " indicate multiplying
Symbol, "/" indicate division arithmetic symbol.
2013, according to the correlation of the application and the reference feature of every group of user, the application and multiple user's are obtained
The correlation of reference feature.
Wherein, the method for obtaining the application and the correlation of the reference feature of multiple user can be with are as follows: according to the application
With the correlation of the reference feature of every group of user, being averaged for the application and the correlation of the reference feature of all grouping users is obtained
The average value or maximum value, are retrieved as the correlation of the application with the reference feature of multiple user by value or maximum value.Certainly,
The application and multiple use can also be obtained using other methods according to the correlation of the application and the reference feature of every group of user
The correlation of the reference feature at family, the present invention is not especially limit this.
It in an alternative embodiment of the invention, can also be by the application and the correlation of the reference feature of all grouping users
Summation is retrieved as the correlation of the application with the reference feature of multiple user；That is, obtain according to the following formula the application with it is multiple
The correlation of the reference feature of user.
Wherein, IV indicates the correlation of the application with the reference feature of multiple user.By by this application and all points
The summation of the correlation of the reference feature of group user, is retrieved as the correlation of the application with the reference feature of multiple user, energy
Enough accuracys for improving accessed correlation, and then can be improved the accuracy of subsequent applications recommendation.
202, multiple application is grouped according to the correlation, obtains first using group, which includes
Multiple correlations are greater than the application of preset threshold.
Wherein, within a preset range, the lower limit of the preset range is that multiple application corresponds to correlation to the preset threshold
Minimum value, the upper limit are the maximum value that multiple application corresponds to correlation, which can be set in the preset range
Any value；For example, the preset threshold can be 0.2.
In embodiments of the present invention, any correlation is greater than the application of the preset threshold, reflects the application and levies with user
Believe that the degree of correlation of feature is higher, divides first into using group by the application that correlation is greater than preset threshold, can be realized and be directed to
First application group is individually established for obtaining the prediction model using recommendation probability, and then can be realized special for user's reference
The high application of sign degree of correlation is recommended, and achievees the purpose that improving application recommends accuracy rate.
In an alternative embodiment of the invention, after being grouped according to the correlation to the multiple application, is also obtained
Two apply group, and the second application group includes the application that multiple correlations are less than or equal to the preset threshold.
Since any correlation is less than or equal to the application of the preset threshold, the application and user's reference feature are reflected
Degree of correlation is lower, therefore by the way that multiple application is divided into one group high with reference feature degree of correlation according to the preset threshold
One group low with reference feature degree of correlation, can be realized, basis is different from reference feature degree of correlation, is every group of application
Corresponding prediction model is established, basis is different from reference feature degree of correlation to reach, and obtains using different prediction models
The recommendation probability of corresponding application, is recommended further according to the recommendation probability, can be further increased using the accuracy rate recommended.
203, group is applied for first, the reference feature based on multiple user establishes corresponding first prediction model, should
First prediction model is used to determine the recommendation probability of application based on reference feature.
In embodiments of the present invention, the method that the reference feature based on multiple user establishes corresponding first prediction model
It can be with are as follows: using the user characteristics of multiple user as training sample, obtain every group using corresponding prediction model by training.
Wherein, which further includes the users such as age, gender Figure Characteristics, online interaction spy other than including reference feature
The features such as sign, history Download History, method used by prediction model of establishing can be SVM (Support Vector
Machine, support vector machines), or the machine learning methods such as maximum entropy or random forest can also be calculated using other
Method establishes prediction model, and the embodiment of the present invention is to the particular user feature as training sample and establishes used by prediction model
Method is not construed as limiting.
It should be noted that for applying group with reference feature correlation higher first, i.e., in the first application group
It is more demanding using the creditworthiness to user, it will include that reference is special when establishing the first application corresponding prediction model of group
The user characteristics of multiple users of sign establish first prediction model as training sample, so that recommending to be somebody's turn to do to any user
First applies in group in application, can recommend be suitble to the user, the i.e. user to the user according to the reference feature of the user
Download the big application of possibility.
By establishing corresponding prediction model to every group of application being grouped according to correlation, can be realized according to different pre-
It surveys model and obtains the recommendation probability of different applications, it, can when carrying out further according to the recommendation probability of the different application using recommending
Improve the accuracy rate that application is recommended.
In an alternative embodiment of the invention, for including application of multiple correlations less than or equal to the preset threshold
Second applies group, establish obtain the method for the second prediction model of the recommendation probability of each application in the second application group can be with
Are as follows: the user characteristics based on multiple user in addition to the reference feature establish second prediction model.
Specifically, for applying group, i.e. application pair in the second application group with reference feature correlation lower second
The creditworthiness requirement of user is lower, will be in addition to reference feature when establishing the first application corresponding prediction model of group
The user characteristics of multiple user establish second prediction model as training sample so as to any user recommend this
Two apply in groups in application, can be recommended according to the other users feature in addition to the reference feature of the user to the user
It is suitble to the application of the user, i.e. the user downloads the big application of possibility.It should be noted that second prediction model can be
Traditional CT R (Click-Through Rate, clicking rate) prediction model.
By establish respectively with reference feature correlation it is higher first application group and with reference feature correlation it is lower
The prediction model of second application group, and when carrying out to user using recommending, recommended not by corresponding prediction model to user
It with using the application in group, can further increase using the accuracy rate recommended, and then can be improved user's downloading and recommend to answer
Probability, to improve the conversion ratio for recommending application.
In yet another embodiment of the invention, in order to further increase using recommend accuracy rate, for this first apply group,
It can also make further packet transaction, to establish every group and apply corresponding sub- prediction model according to the group result being grouped again,
To obtain the recommendation probability of each application in every group of application.It may include following to first method for applying group to be further grouped
Two kinds:
The first, the correlation applied in the first application group is divided into multiple sections by default size；By this first
Same application group is divided into using the application that correlation in group belongs to same section；According to group result, based on multiple user's
Reference feature is that corresponding sub- prediction model, the sub- prediction model of the corresponding application group in the bigger section of correlation are established in every group of application
Reference feature weight it is bigger.
Wherein, which can be determined as any fixed value, can also be according to point of correlation in the first application group
Cloth determines, is such as determined according to minimum relatedness and maximum correlation, and the 1/3 of maximum correlation and minimum relatedness difference is determined
Size is preset for this, which is preset into size according to this and is further divided into three using group, it is of course also possible to use other
Method determines the default size, and the present invention is not especially limit this.
For example, working as maximum correlation in the first application group is 0.9, when minimum relatedness is 0.6, which is 0.1,
Size is preset according to this, the correlation applied in the first application group is divided into three groups: [0.6,0.7), [0.7,0.8),
[0.8,0.9], this first application group in, correlation size [0.6,0.7) section application be classified as one group, correlation size exists
[0.7,0.8) application in section is classified as one group, and application of the correlation size in [0.8,0.9] section is classified as one group.
It, can be according to application and user's reference by making further packet transaction to the first application group according to the correlation
The degree of correlation of feature is different, establishes the different prediction model of reference feature weight, obtains every group of application after being grouped again
Recommend probability, can achieve the purpose that further increase the accuracy using recommending.
It second, can also be according to using class for applying the application in group with reference feature correlation higher first
Type does further grouping to the application in first application, establishes corresponding prediction model, specific method further according to group result
It can be with are as follows: by application type in the first application group be that the application of specified type is classified as one group, obtain third application group；By this
Application type is that the other kinds of application in addition to the specified type is classified as one group in one application group, obtains the 4th using group；
Based on the third application group, the user characteristics based on multiple user establish the first sub- prediction model；Group is applied based on the 4th,
User characteristics based on multiple user establish the second sub- prediction model；Wherein, the reference feature in the first sub- prediction model
Weight is greater than the reference feature weight in the second sub- prediction model.
Wherein, which can be determined as any kind in the types such as online shopping class, financial class by developer, should
Specified type is also possible to require to determine according to the different of user credit degree, for example, by the specified type be determined as to
Family creditworthiness requires the type of higher application.It is, of course, also possible to determine that the specified type, the present invention are real using other methods
Example is applied to the concrete application type of specified type and determines that method is not construed as limiting.
For example, being that the application of financial class is returned by application type in the first application group when the specified type is financial class
For third application group, it is that the application of non-financial class is classified as the 4th using group by application type in the first application group, is established
First sub- prediction model is used to obtain the recommendation probability of financial class application, and the second sub- prediction model is for obtaining non-financial class application
Recommendation probability.
It should be noted that being applied due to the finance class is the application more demanding to user credit degree, for example, financing
Using, equity investment application etc., financial class application is obtained by using the first bigger sub- prediction model of reference feature weight
Recommend probability, improves user credit degree to the influence of probability is recommended, further increase to reach using recommendation accuracy rate
Purpose.And relatively low application is required user credit degree for other kinds of, for example, social category application, online shopping class
Using etc., the recommendation for obtaining non-financial class application by using the sub- prediction model of reference feature weight relatively small second is general
Rate reduces influence of the user credit degree to probability is recommended, and can also achieve the purpose that improving application recommends accuracy rate.
By using the different prediction model of reference feature weight, the recommendation probability applied to different type is obtained, especially
Application for specified type, by improving the weight of reference feature, when obtaining the recommendation probability of the type application, Neng Goujin
One step improves the accuracy rate for recommending application, so that the application recommended more meets the downloading demand of user, so as to
Achieve the purpose that further increase using conversion ratio.
The process for making further packet transaction to the first application group, can be using any side in above two method
Method is realized, can also make further packet transaction to the first application group using other methods, the embodiment of the present invention does not make this
It is specific to limit.
Above-mentioned steps 201 to step 203 is for the application and/or different types of application foundation correspondence in different grouping
Prediction model process, when simultaneously be grouped according to correlation and application type when, to after grouping application establish correspond to
The process of prediction model can be indicated with Fig. 2 B.Specifically, the reference of each application and multiple users in multiple applications is obtained
The correlation of feature is grouped multiple application according to correlation and application type, and correlation is less than or equal to default threshold
The application of value is classified as one group, obtains second using group, and correlation is greatly with preset threshold and application type is the application of specified type
It is classified as one group, obtains third application group, correlation is greater than the preset threshold and application type and is classified as the application of non-designated type
One group, the 4th is obtained using group.
Since the second application group and the degree of correlation of reference feature are minimum, it can be by traditional CTR prediction model
As second prediction model；The the first sub- prediction model established for third application group can be used for obtaining the third application
In any application recommendation probability, can also individually establish for each application in third application the group and be pushed away for obtaining to correspond to
The prediction model of probability is recommended, the reference feature weight of each prediction model is different, and the embodiment of the present invention is not construed as limiting this；For
4th applies group, establishes the second sub- prediction model according to the user characteristics including reference feature, the second son prediction mould
Type can be logistic regression prediction model, the feature of the reference feature weight of the second sub- prediction model less than the first sub- prediction model
Minimal characteristic weight in weight, or all prediction models corresponding less than third application group.
204, the reference feature based on user to be recommended and first prediction model are recommended to answer to the user to be recommended
With.
When receiving the recommendation application acquisition request that the user to be recommended sends, server obtains the user's to be recommended
User characteristics, and in the first prediction model that the user characteristics input step 203 of the user to be recommended is established, with obtain this
The recommendation probability of different application, according to the recommendation probability selection of the different application pushes away to the user to be recommended in one application group
Recommend the application in the first application group.For example, application of the probability greater than 0.5 will be recommended to recommend the user to be recommended, at this
The terminal of user to be recommended shows the recommendation application, the display mode of recommendation application can according to Apply Names initial it is suitable
Sequence is shown, or according to recommending the sequence of probability from big to small to show, can also be shown using other methods, present example pair
The method of being particularly shown is not construed as limiting.
For example, when terminal detects that user to be recommended opens application and recommends the page or on application service platform to recommendation
When the trigger action of application option, terminal to server, which is sent, to be recommended to apply acquisition request, which can be with
The user characteristics of the user to be recommended are carried, or carry the user identifier of the user to be recommended, so that server being capable of root
The user characteristics of the user to be recommended are inquired according to the user identifier of the user to be recommended, and then according to the use of the user to be recommended
Family feature recommends application to the user to be recommended.
The recommendation probability of each application in the first application group is obtained by the first prediction model, and according to the recommendation probability
Recommend the application in the first application group to user, can be improved the accuracy rate recommended the application in the first application group.
In an alternative embodiment of the invention, after for being grouped according to correlation to multiple application, obtain first
Using group, specific method for being recommended to the application in the first application group can be with are as follows: the user of the user to be recommended is special
Sign inputs first prediction model, obtains the recommendation probability of each application in the first application group, according in the first application group
The recommendation probability of each application is recommended to recommend probability to be greater than answering for predetermined probabilities in the first application group to the user to be recommended
With.
The predetermined probabilities can be set to it is any be greater than 0 numerical value less than 1, can also be obtained according to the user to be recommended selection
The number of the recommendation application taken determines, for example, recommending when user's selection obtains 50 in application, the predetermined probabilities are determined as
0.7, when user selects to obtain 100 recommendation probability, which is determined as 0.5；It is of course also possible to use its other party
Method determines the predetermined probabilities, and the present invention is not especially limit this.
The application in the first application group is recommended by the above method, can according to the difference of the predetermined probabilities, to
The user to be recommended recommends the application of different number, so as to better meet the recommended requirements of user.
In yet another embodiment of the invention, after being grouped according to correlation to multiple application, obtain second
Using group, specific method for being recommended to the application in the second application group can be with are as follows: will remove the reference of the user to be recommended
The user characteristics of the user to be recommended other than feature input the second class prediction model, obtain each answering in the second application group
Recommendation probability recommends this second to answer according to the recommendation probability of each application in the second application group to the user to be recommended
With the application for recommending probability to be greater than the predetermined probabilities in group.
In embodiments of the present invention, the application in the first application group only can be recommended to user, it is less to recommend to user
Quantity, with the higher application of reference feature degree of correlation, allow users to more clearly to check and download answering of being recommended
With；Further, when receiving user and obtaining the request for more recommending application, then answering into user's recommendation the second application group
With to meet the needs of user obtains more recommendation applications；It is of course also possible to when receiving recommendation using acquisition request, together
When to user recommend the application in the first application group and the second application group, by corresponding the recommendations application of different application group using paging
The method of display is shown, to distinguish the application recommended using different prediction models, while so that the page is cleaner and tidier, Xiang Yong
Family shows that the page is recommended in the bigger application of information content.
In an alternative embodiment of the invention, recommend the page that can also show specified function choosing-item in the application, the specified function
Energy option is used to obtain the recommendation application in the first application group, and the display area attachment of the specified function choosing-item can be with display reminding
Information, the prompt information are special with user's reference for prompting user to trigger recommendation accessed by the specified function choosing-item and apply
Levy the big application of correlation, such as financial class, the application of online shopping class.When terminal detects that user specifies the triggering of function choosing-item to this
When operation, is sent to server and obtain the specified request for recommending application, so that server is pushed away according to the request using the first
It recommends method and recommends to apply to user.
Recommend to apply by the above method, can be improved using the specific aim recommended, and then can recommend more to accord with to user
The application of user demand is closed, to achieve the purpose that improve using conversion ratio.
By according to the preset threshold by multiple application be divided into reference feature degree of correlation it is high first application group and
Second low with reference feature degree of correlation applies group, obtains in every group of application often according to every group using corresponding prediction model
The recommendation probability of a application, and carried out can be improved the success rate using recommending using recommendation according to the recommendation probability.
It should be noted that for first making after further packet transaction to be grouped again as a result, obtaining using group to this
Afterwards in every group of application the recommendation probability of each application method, and it is every in the above-mentioned acquisition first application group and the second application group
The method of the recommendation probability of a application similarly, does not repeat herein.
Recommended method is applied provided by the embodiment of the present invention, by obtaining for indicating using the reference feature with user
Degree of correlation correlation, and according in multiple applications it is each application and the reference feature of multiple users correlation, to
The reference feature degree of correlation of user it is high first apply group, using based on multiple users reference feature establish first prediction
Model obtains the recommendation probability of each application in the first application group, can be improved using recommendation according to the recommendation probability
To the accuracy rate in the first application group using recommendation, so that the application recommended more meets the downloading demand of user, so as to
Enough conversion ratios for improving application；By using the different prediction model of reference feature weight, acquisition pushes away different type application
Probability is recommended, the recommendation accuracy rate to the application different from reference feature degree of correlation can be further increased.
Fig. 3 is a kind of application recommendation apparatus block diagram provided in an embodiment of the present invention.Referring to Fig. 3, which includes software phase
Closing property obtains module 301, grouping module 302, model building module 303 and recommending module 304.
Correlation obtains module 301, for obtaining the phase of each application and the reference feature of multiple users in multiple applications
Guan Xing；
Grouping module 302, the correlation for being got according to correlation acquisition module 301 is to the multiple
Using being grouped, first is obtained using group, the first application group includes the application that multiple correlations are greater than preset threshold；
Model building module 303, first for obtaining for the grouping module 302 applies group, based on the multiple
The reference feature of user establishes corresponding first prediction model, and first prediction model is used to be determined based on reference feature and be applied
Recommendation probability；
Recommending module 304, the institute for reference feature and the foundation of the model building module 303 based on user to be recommended
The first prediction model is stated, recommends application to the user to be recommended.
In the first possible implementation provided by the invention, the recommending module 304 is used for:
The reference feature of the user to be recommended is inputted into first prediction model, is obtained every in the first application group
The recommendation probability of a application is recommended according to the recommendation probability of each application in the first application group to the user to be recommended
Probability is recommended to be greater than the application of predetermined probabilities in the first application group.
In second provided by the invention possible implementation, the model building module 303 is also used to:
Group, the use based on the multiple user in addition to the reference feature are applied for second obtained by grouping
Family feature establishes the second prediction model, and second prediction model is used for true based on the user characteristics in addition to the reference feature
Surely the recommendation probability applied, the second application group include the application that multiple correlations are less than or equal to the preset threshold.
In the third possible implementation provided by the invention, the recommending module 304 is used for:
By the user characteristics of the user to be recommended in addition to the reference feature of the user to be recommended input described the
Two prediction models obtain the recommendation probability of each application in the second application group, each answer according in the second application group
Recommendation probability is recommended to recommend probability to be greater than answering for the predetermined probabilities in the second application group to the user to be recommended
With.
In the 4th kind of possible implementation provided by the invention, the model building module 303 is used for:
It is that the application of specified type is classified as one group by application type in the first application group, obtains third application group；
It is that the other kinds of application in addition to the specified type is classified as one by application type in the first application group
Group obtains the 4th using group；
Based on the third application group, the user characteristics based on the multiple user establish the first sub- prediction model；
Group is applied based on the described 4th, the user characteristics based on the multiple user establish the second sub- prediction model；
Wherein, the reference feature weight in the described first sub- prediction model is greater than the reference in the described second sub- prediction model
Feature weight.
In the 5th kind of possible implementation provided by the invention, the model building module 303 is also used to:
The correlation applied in the first application group is divided into multiple sections by default size；
Divide the application that correlation belongs to same section in the first application group into same application group；
According to group result, the reference feature based on the multiple user is that corresponding sub- prediction mould is established in every group of application
The reference feature weight of type, the sub- prediction model of the corresponding application group in the bigger section of correlation is bigger.
In the 6th kind of possible implementation provided by the invention, the correlation obtains module 301 and is used for:
For any application in the multiple application, according to the reference feature of the multiple user, to the multiple use
Family is grouped；
The downloading situation of the application and the multiple user obtain the downloading situation of the application according to every group of user
Take the correlation of the application and the reference feature of every group of user；
According to the correlation of the application and the reference feature of every group of user, the application is obtained with the multiple user's
The correlation of reference feature.
In the 7th kind of possible implementation provided by the invention, the correlation obtains module 301 and is used for:
The correlation of the application and the reference feature of every group of user is obtained according to the following formula:
Wherein, i=1,2 ..., n, n indicate the user group number obtained after being grouped to the multiple user, IViIt indicates
The correlation of the application and the reference feature of i-th group of user, GiIndicate the user that the application is downloaded in i-th group of user
Quantity, BiIndicate the number of users for not downloading the application in i-th group of user, GTIt indicates to download institute in the multiple user
State the number of users of application, BTIndicate the number of users for not downloading the application in the multiple user.
In the 8th kind of possible implementation provided by the invention, the correlation obtains module 301 for according to the following formula
Obtain the correlation of the application and the reference feature of the multiple user:
Wherein, IV indicates the correlation of the application and the reference feature of the multiple user.
It should be understood that application recommendation apparatus provided by the above embodiment is being recommended in application, only with above-mentioned each function
The division progress of module can according to need and for example, in practical application by above-mentioned function distribution by different function moulds
Block is completed, i.e., the internal structure of equipment is divided into different functional modules, to complete all or part of function described above
Energy.In addition, application recommendation apparatus provided by the above embodiment belongs to same design with using recommended method embodiment, it is specific real
Existing process is detailed in embodiment of the method, and which is not described herein again.
Fig. 4 is a kind of structural schematic diagram of device 400 provided in an embodiment of the present invention.For example, device 400 can be provided
For a server.Referring to Fig. 4, it further comprises one or more processors, Yi Jiyou that device 400, which includes processing component 422,
Memory resource representated by memory 432, can be by the instruction of the execution of processing component 422, such as application program for storing.
The application program stored in memory 432 may include it is one or more each correspond to one group of instruction module.
In addition, processing component 422 is configured as executing instruction, it is above-mentioned using recommended method to execute.
Device 400 can also include the power management that a power supply module 426 is configured as executive device 400, and one has
Line or radio network interface 450 are configured as device 400 being connected to network and input and output (I/O) interface 458.Dress
Setting 400 can operate based on the operating system for being stored in memory 432, such as Windows ServerTM, Mac OS XTM,
UnixTM,LinuxTM, FreeBSDTMOr it is similar.
Those of ordinary skill in the art will appreciate that realizing that all or part of the steps of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can store in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (18)
1. a kind of apply recommended method, which is characterized in that the described method includes:
Obtain the correlation of each application and the reference feature of multiple users in multiple applications；
The multiple application is grouped according to the correlation, first is obtained and applies group, include in the first application group
Multiple correlations are greater than the application of preset threshold；
Group is applied for first, the reference feature based on the multiple user establishes corresponding first prediction model, and described first
Prediction model is used to determine the recommendation probability of application based on reference feature；
Reference feature and first prediction model based on user to be recommended recommend application to the user to be recommended.
2. the method according to claim 1, wherein the reference feature and described based on user to be recommended
One prediction model recommends application to the user to be recommended, comprising:
The reference feature of the user to be recommended is inputted into first prediction model, obtains each answering in the first application group
Recommendation probability, according to the recommendation probability of each application in the first application group, to described in user's recommendation to be recommended
Probability is recommended to be greater than the application of predetermined probabilities in first application group.
3. the method according to claim 1, wherein described carry out the multiple application according to the correlation
After grouping, the method also includes:
Group is applied for second obtained by grouping, the user based on the multiple user in addition to the reference feature is special
Sign establishes the second prediction model, and second prediction model is used to answer based on the user characteristics determination in addition to the reference feature
Recommendation probability, the second application group include the application that multiple correlations are less than or equal to the preset threshold.
4. according to the method described in claim 3, it is characterized in that, the reference feature and described based on user to be recommended
One prediction model recommends application to the user to be recommended, comprising:
The user characteristics input described second of the user to be recommended in addition to the reference feature of the user to be recommended is pre-
Model is surveyed, the recommendation probability of each application in the second application group is obtained, according to each application in the second application group
Recommend probability, recommends the application for recommending probability to be greater than predetermined probabilities in the second application group to the user to be recommended.
5. being based on the multiple user the method according to claim 1, wherein described apply group for first
Reference feature establish corresponding first prediction model, comprising:
It is that the application of specified type is classified as one group by application type in the first application group, obtains third application group；
It is that the other kinds of application in addition to the specified type is classified as one group by application type in the first application group, obtains
Group is applied to the 4th；
Based on the third application group, the user characteristics based on the multiple user establish the first sub- prediction model；
Group is applied based on the described 4th, the user characteristics based on the multiple user establish the second sub- prediction model；
Wherein, the reference feature weight in the described first sub- prediction model is greater than the reference feature in the described second sub- prediction model
Weight.
6. being based on the multiple user the method according to claim 1, wherein described apply group for first
Reference feature establish corresponding first prediction model, comprising:
The correlation applied in the first application group is divided into multiple sections by default size；
Divide the application that correlation belongs to same section in the first application group into same application group；
According to group result, the reference feature based on the multiple user is that corresponding sub- prediction model, phase are established in every group of application
The reference feature weight of the sub- prediction model of the corresponding application group in Guan Xingyue big section is bigger.
7. the method according to claim 1, wherein described obtain each application and multiple users in multiple applications
Reference feature correlation, comprising:
For any application in the multiple application, according to the reference feature of the multiple user, to the multiple user into
Row grouping；
Institute is obtained to the downloading situation of the application to the downloading situation of the application and the multiple user according to every group of user
It states using the correlation with the reference feature of every group of user；
According to the correlation of the application and the reference feature of every group of user, the reference of the application and the multiple user is obtained
The correlation of feature.
8. the method according to the description of claim 7 is characterized in that it is described according to every group of user to the downloading situation of the application
With the multiple user to the downloading situation of the application, it is related to the reference feature of every group of user to obtain the application
Property, comprising:
The correlation of the application and the reference feature of every group of user is obtained according to the following formula:
Wherein, i=1,2 ..., n, n indicate the user group number obtained after being grouped to the multiple user, IViIt is answered described in expression
With the correlation of the reference feature with i-th group of user, GiIndicate the number of users that the application is downloaded in i-th group of user, Bi
Indicate the number of users for not downloading the application in i-th group of user, GTIt indicates to download the application in the multiple user
Number of users, BTIndicate the number of users for not downloading the application in the multiple user.
9. according to the method described in claim 8, it is characterized in that, obtaining the application according to the following formula with the multiple user's
The correlation of reference feature:
Wherein, IV indicates the correlation of the application and the reference feature of the multiple user.
10. a kind of apply recommendation apparatus, which is characterized in that described device includes:
Correlation obtains module, for obtaining the correlation of each application and the reference feature of multiple users in multiple applications；
Grouping module divides the multiple application for obtaining the correlation that module is got according to the correlation
Group obtains first using group, and the first application group includes the application that multiple correlations are greater than preset threshold；
Model building module, first for obtaining for the grouping module applies group, the reference based on the multiple user
Feature establishes corresponding first prediction model, and first prediction model is used to determine that the recommendation of application is general based on reference feature
Rate；
Recommending module, first prediction for reference feature and model building module foundation based on user to be recommended
Model recommends application to the user to be recommended.
11. device according to claim 10, which is characterized in that the recommending module is used for:
The reference feature of the user to be recommended is inputted into first prediction model, obtains each answering in the first application group
Recommendation probability, according to the recommendation probability of each application in the first application group, to described in user's recommendation to be recommended
Probability is recommended to be greater than the application of predetermined probabilities in first application group.
12. device according to claim 10, which is characterized in that the model building module is also used to:
Group is applied for second obtained by grouping, the user based on the multiple user in addition to the reference feature is special
Sign establishes the second prediction model, and second prediction model is used to answer based on the user characteristics determination in addition to the reference feature
Recommendation probability, the second application group include the application that multiple correlations are less than or equal to the preset threshold.
13. device according to claim 12, which is characterized in that the recommending module is used for:
The user characteristics input described second of the user to be recommended in addition to the reference feature of the user to be recommended is pre-
Model is surveyed, the recommendation probability of each application in the second application group is obtained, according to each application in the second application group
Recommend probability, recommends the application for recommending probability to be greater than predetermined probabilities in the second application group to the user to be recommended.
14. device according to claim 10, which is characterized in that the model building module is used for:
It is that the application of specified type is classified as one group by application type in the first application group, obtains third application group；
It is that the other kinds of application in addition to the specified type is classified as one group by application type in the first application group, obtains
Group is applied to the 4th；
Based on the third application group, the user characteristics based on the multiple user establish the first sub- prediction model；
Group is applied based on the described 4th, the user characteristics based on the multiple user establish the second sub- prediction model；
Wherein, the reference feature weight in the described first sub- prediction model is greater than the reference feature in the described second sub- prediction model
Weight.
15. device according to claim 10, which is characterized in that the model building module is also used to:
The correlation applied in the first application group is divided into multiple sections by default size；
Divide the application that correlation belongs to same section in the first application group into same application group；
According to group result, the reference feature based on the multiple user is that corresponding sub- prediction model, phase are established in every group of application
The reference feature weight of the sub- prediction model of the corresponding application group in Guan Xingyue big section is bigger.
16. device according to claim 10, which is characterized in that the correlation obtains module and is used for:
For any application in the multiple application, according to the reference feature of the multiple user, to the multiple user into
Row grouping；
Institute is obtained to the downloading situation of the application to the downloading situation of the application and the multiple user according to every group of user
It states using the correlation with the reference feature of every group of user；
According to the correlation of the application and the reference feature of every group of user, the reference of the application and the multiple user is obtained
The correlation of feature.
17. device according to claim 16, which is characterized in that the correlation obtains module and is used for:
The correlation of the application and the reference feature of every group of user is obtained according to the following formula:
Wherein, i=1,2 ..., n, n indicate the user group number obtained after being grouped to the multiple user, IViIt is answered described in expression
With the correlation of the reference feature with i-th group of user, GiIndicate the number of users that the application is downloaded in i-th group of user, Bi
Indicate the number of users for not downloading the application in i-th group of user, GTIt indicates to download the application in the multiple user
Number of users, BTIndicate the number of users for not downloading the application in the multiple user.
18. device according to claim 17, which is characterized in that the correlation obtains module for obtaining according to the following formula
The correlation of the application and the reference feature of the multiple user:
Wherein, IV indicates the correlation of the application and the reference feature of the multiple user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610814453.7A CN107807935B (en) | 2016-09-09 | 2016-09-09 | Using recommended method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610814453.7A CN107807935B (en) | 2016-09-09 | 2016-09-09 | Using recommended method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107807935A CN107807935A (en) | 2018-03-16 |
| CN107807935B true CN107807935B (en) | 2019-05-07 |
Family
ID=61569668
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610814453.7A Active CN107807935B (en) | 2016-09-09 | 2016-09-09 | Using recommended method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107807935B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108664286B (en) * | 2018-05-10 | 2021-07-09 | Oppo广东移动通信有限公司 | Application program preloading method and device, storage medium and mobile terminal |
| CN108762844B (en) * | 2018-05-29 | 2021-09-07 | Oppo广东移动通信有限公司 | Application program preloading method and device, storage medium and terminal |
| CN109151019A (en) * | 2018-08-20 | 2019-01-04 | 彩讯科技股份有限公司 | A kind of application method for down loading, device, equipment and storage medium |
| CN109299349B (en) * | 2018-09-11 | 2021-02-19 | 广州视源电子科技股份有限公司 | Application recommendation method and device, equipment and computer-readable storage medium |
| CN115049039B (en) * | 2021-03-08 | 2023-11-14 | 北京金茂绿建科技有限公司 | Neural network-based state identification method, neural network training method and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103034508A (en) * | 2011-10-10 | 2013-04-10 | 腾讯科技（深圳）有限公司 | Software recommending method and software recommending system |
| CN103338223A (en) * | 2013-05-27 | 2013-10-02 | 清华大学 | Recommendation method, client and server of mobile application |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080010198A1 (en) * | 2000-09-07 | 2008-01-10 | Lsq Ii, Llc | System for and method of providing financial and transaction management services over a network |
| US20090018955A1 (en) * | 2007-07-13 | 2009-01-15 | Yen-Fu Chen | Method and apparatus for providing user access to payment methods |
-
2016
- 2016-09-09 CN CN201610814453.7A patent/CN107807935B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103034508A (en) * | 2011-10-10 | 2013-04-10 | 腾讯科技（深圳）有限公司 | Software recommending method and software recommending system |
| CN103338223A (en) * | 2013-05-27 | 2013-10-02 | 清华大学 | Recommendation method, client and server of mobile application |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107807935A (en) | 2018-03-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107807935B (en) | Using recommended method and device | |
| CN107741976B (en) | Intelligent response method, device, medium and electronic equipment | |
| US10650818B2 (en) | Semi-supervised question answering machine | |
| JP2021519472A (en) | Knowledge sharing method, dialogue method, knowledge sharing device, dialogue device, electronic device and storage medium between dialogue systems | |
| CN104156246B (en) | The method and device of gray scale upgrading | |
| CN110009486B (en) | Method, system, equipment and computer readable storage medium for fraud detection | |
| US11748452B2 (en) | Method for data processing by performing different non-linear combination processing | |
| CN106708858A (en) | Information recommendation method and device | |
| CN110413867B (en) | Method and system for content recommendation | |
| Mei et al. | An adaptive service selection approach to service composition | |
| CN105303410B (en) | Virtual article sharing method and device | |
| CN106534281A (en) | Data request responding method, apparatus and system | |
| CN107832941A (en) | Order processing method and device | |
| CN111008335A (en) | Information processing method, device, equipment and storage medium | |
| US20240005165A1 (en) | Machine learning model training method, prediction method therefor, apparatus, device, computer-readable storage medium, and computer program product | |
| CN107562847A (en) | Information processing method and related product | |
| CN109934631A (en) | Question and answer information processing method, device and computer equipment | |
| CN112488163A (en) | Abnormal account identification method and device, computer equipment and storage medium | |
| CN111008873A (en) | User determination method and device, electronic equipment and storage medium | |
| CN118349812A (en) | Model determination method, apparatus, device, storage medium, and program product | |
| WO2024152686A1 (en) | Method and apparatus for determining recommendation index of resource information, device, storage medium and computer program product | |
| CN113138702A (en) | Information processing method, information processing device, electronic equipment and storage medium | |
| CN114170004A (en) | Scoring decision-making method, device, equipment and storage medium based on multiple events | |
| CN114327693A (en) | AB experiment method and system, terminal device and storage medium | |
| CN114529143A (en) | Method and device for recommending outbound clues and electronic equipment |
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 |