CN107977445A - Application program recommends method and device - Google Patents

Application program recommends method and device Download PDF

Info

Publication number
CN107977445A
CN107977445A CN201711307219.6A CN201711307219A CN107977445A CN 107977445 A CN107977445 A CN 107977445A CN 201711307219 A CN201711307219 A CN 201711307219A CN 107977445 A CN107977445 A CN 107977445A
Authority
CN
China
Prior art keywords
application
application program
user
list
country
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
Application number
CN201711307219.6A
Other languages
Chinese (zh)
Other versions
CN107977445B (en
Inventor
蔡馥励
王长路
庞国盛
李涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Kylin Hesheng Network Technology Co Ltd
Original Assignee
Beijing Kylin Hesheng Network Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Kylin Hesheng Network Technology Co Ltd filed Critical Beijing Kylin Hesheng Network Technology Co Ltd
Priority to CN201711307219.6A priority Critical patent/CN107977445B/en
Publication of CN107977445A publication Critical patent/CN107977445A/en
Application granted granted Critical
Publication of CN107977445B publication Critical patent/CN107977445B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2462—Approximate or statistical queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • 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)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Probability & Statistics with Applications (AREA)
  • Fuzzy Systems (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Stored Programmes (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The embodiment of the present application provides a kind of application program and recommends method and device, and wherein method includes：According to the relevant information of each user, multiple users are divided into multiple user groups；When the application searches for receiving targeted customer are asked, whether the number of users of user group is more than predetermined quantity where judging targeted customer；If more than, it is determined that the application searches of targeted customer ask the category of employment of corresponding searched application；If category of employment is life kind application, application program is then recommended to targeted customer according to the corresponding first application program affinity list of country that is currently located of targeted customer, such as searched application is non-life kind application, then recommends application program to targeted customer according to the corresponding second application program affinity list of place user group of targeted customer.Pass through the embodiment of the present application, it is possible to increase the recommendation precision of application program.

Description

Application program recommends method and device
Technical field
This application involves application program to recommend field, more particularly to a kind of application program to recommend method and device.
Background technology
With the popularization of intelligent terminal and the development of mobile terminal technology, more and more user's selection answering in mobile terminal With downloading required application software in shop.According to statistics, the number of applications in shop is applied for Android system, GooglePlay 3,000,000 are alreadyd exceed, user finds required application in magnanimity application will expend great effort, and therefore, user is answered Come into being with the technology precisely recommended.
At present, user is carried out using the mode recommended being mainly that the download respectively applied in application shop is united Meter sequence, generate total download list and the download list of application of all categories, according to total download list and it is of all categories should Download list, recommends application program for user.
However, in aforesaid way, application program is recommended based on download list, the application program recommended for each user is equal Identical, not in view of the personalized difference between different user, therefore the recommendation precision of application program is poor.
The content of the invention
The purpose of the embodiment of the present application is to provide a kind of application program and recommends method and device, to improve pushing away for application program Recommend precision.
In order to solve the above technical problems, what the embodiment of the present application was realized in：
In a first aspect, the embodiment of the present application, which provides a kind of application program, recommends method, including：
According to the corresponding relevant information of each user, the whole network user is divided into multiple user groups；Wherein, the phase Closing information includes religious belief, place city, the country one belongs to, language used, colonization country, colonization city, the place city Income level, it is described settle down city income level at least one of；
When the application searches for receiving targeted customer are asked, the number of users of user group where judging the targeted customer Whether predetermined quantity is more than；
If more than predetermined quantity, it is determined that the application searches of the targeted customer ask the industry of corresponding searched application Classification；
Such as category of employment is life kind application, then is currently located country corresponding first according to the targeted customer Application program affinity list recommends application program to the targeted customer, and such as category of employment is non-life kind application, then root According to the corresponding second application program affinity list of the place user group of the targeted customer journey is applied to targeted customer recommendation Sequence.
Second aspect, the embodiment of the present application provide a kind of application program recommendation apparatus, including：
User clustering module, for according to the corresponding relevant information of each user, the whole network user being divided into multiple User group；Wherein, the relevant information includes religious belief, place city, the country one belongs to, language used, colonization country, colonization At least one of in city, the income level in the place city, the income level for settling down city；
Quantity discrimination module, for when receiving the application searches request of targeted customer, judging the targeted customer institute Whether it is more than predetermined quantity in the number of users of user group；
Classification discrimination module, for if more than predetermined quantity, it is determined that the application searches request of the targeted customer corresponds to Searched application category of employment；
It is life kind application for such as described category of employment, then according to the current of the targeted customer using recommending module The corresponding first application program affinity list of the country one belongs to recommends application program to the targeted customer, and such as category of employment is Non- life kind application, then according to the corresponding second application program affinity list of place user group of the targeted customer to the mesh Mark user and recommend application program.
The third aspect, the embodiment of the present application provide a kind of application program recommendation apparatus, including：Memory, processor and The computer program that can be run on the memory and on the processor is stored in, the computer program is by the processing The step of application program as described in above-mentioned first aspect recommends method is realized when device performs.
Fourth aspect, the embodiment of the present application provide a kind of storage medium, computer journey are stored with the storage medium Sequence, realizes that the application program as described in above-mentioned first aspect recommends the step of method when the computer program is executed by processor Suddenly.
By the method and device in the embodiment of the present application, multiple users can be drawn according to a variety of relevant informations of user It is divided into multiple user groups, and life kind application program is searched in user, and the number of users of the place user group of user is more than in advance During fixed number amount, application program is recommended to user according to the corresponding first application program affinity list of country that is currently located of user, Also, non-life kind application program is searched in user, and when the number of users of the place user group of user is more than predetermined quantity, root Recommend application program to user according to the corresponding second application program affinity list of the place user group of user, so that based on user's Searching request, the place user group of user and the country one belongs to of user, personalizedly recommend application program to user, improve application The recommendation precision of program.Also, life kind is searched in user in application, recommending to apply to user based on the country one belongs to of user Program, the application program for enabling to recommend meets the habits and customs of the country one belongs to of user, so that the application for improving user is searched Cable body is tested.Also, in the non-life kind of user's search in application, the place user group based on user recommends application program to user, Enable to the place user group of the application program and user recommended to match, i.e., the classification situation with user matches, so that Improve the application searches experience of user.
Brief description of the drawings
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, below will be to embodiment or existing There is attached drawing needed in technology description to be briefly described, it should be apparent that, drawings in the following description are only this Some embodiments described in application, for those of ordinary skill in the art, in the premise of not making the creative labor property Under, other attached drawings can also be obtained according to these attached drawings.
Fig. 1 is the flow diagram that the application program that one embodiment of the application provides recommends method；
Fig. 2 is the flow diagram that the application program that another embodiment of the application provides recommends method；
Fig. 3 is the module composition schematic diagram for the application program recommendation apparatus that one embodiment of the application provides；
Fig. 4 is the module composition schematic diagram for the application program recommendation apparatus that another embodiment of the application provides；
Fig. 5 is the structure diagram for the application program recommendation apparatus that one embodiment of the application provides.
Embodiment
It is in order to make those skilled in the art better understand the technical solutions in the application, real below in conjunction with the application The attached drawing in example is applied, the technical solution in the embodiment of the present application is clearly and completely described, it is clear that described implementation Example is merely a part but not all of the embodiments of the present application.It is common based on the embodiment in the application, this area Technical staff's all other embodiments obtained without creative efforts, should all belong to the application protection Scope.
The embodiment of the present application provides a kind of application program and recommends method, a kind of application program recommendation apparatus, a kind of application Program recommendation apparatus and a kind of storage medium, wherein, application program recommends method to apply in mobile terminal side, by mobile whole End performs, additionally it is possible to applies in server side, is performed by server, which can be smart mobile phone, computer, tablet electricity The equipment such as brain.
Fig. 1 is the flow diagram that the application program that one embodiment of the application provides recommends method, as shown in Figure 1, the party Method includes：
Step S102, according to the corresponding relevant information of each user, multiple user groups are divided into by the whole network user, its In, relevant information includes religious belief, place city, the country one belongs to, language used, colonization country, colonization city, place city Income level, settle down city income level at least one of；The whole network user refers to whole users in internet；
Step S104, when the application searches for receiving targeted customer are asked, the use of user group where judging targeted customer Whether amount amount is more than predetermined quantity；
Step S106, if more than predetermined quantity, it is determined that the application searches of targeted customer ask corresponding searched application Category of employment；
Step S108, if category of employment is life kind application, is then currently located country corresponding the according to targeted customer One application program affinity list recommends application program to targeted customer, if category of employment is non-life kind application, then according to target The corresponding second application program affinity list of place user group of user recommends application program to targeted customer.
By the embodiment of the present application, multiple users can be divided into multiple users according to a variety of relevant informations of user Group, and life kind application program is searched in user, and when the number of users of the place user group of user is more than predetermined quantity, according to The corresponding first application program affinity list of country that is currently located of user recommends application program to user, also, is searched in user Suo Fei life kind application programs, and when the number of users of the place user group of user is more than predetermined quantity, according to the place of user The corresponding second application program affinity list of user group recommends application program to user, so that the searching request based on user, use The place user group at family and the country one belongs to of user, personalizedly recommend application program to user, improve the recommendation of application program Precision.Also, life kind is searched in user in application, recommending application program to user based on the country one belongs to of user, can So that the application program recommended meets the habits and customs of the country one belongs to of user, so as to improve the application searches experience of user.And And non-life kind is searched in application, the place user group based on user enables to push away to user's recommendation application program in user The application program and the place user group of user recommended match, i.e., the classification situation with user matches, so as to improve user's Application searches are experienced.
In the present embodiment, it is contemplated that religious belief, place city, the country one belongs to, language used, colonization country, colonization city City, the income level in place city, the income level in colonization city can represent the regional characteristic of user, therefore set above-mentioned phase Closing information includes at least one of above- mentioned information, wherein, the regional characteristic of user is exemplified as, and is settled down in a line city of China City, tier 2 cities of the tabernacle at present in the U.S..
In above-mentioned steps S102, according to the corresponding relevant information of each user, the whole network user is divided into multiple use Family group, specifically, according to default clustering algorithm, according to the corresponding relevant information of each user, carries out the whole network user Cluster, obtains multiple user groups.
Specifically, default clustering algorithm can be k-means clustering algorithms, according to k-means clustering algorithms, according to every The corresponding relevant information of a user, to the whole network, user clusters, and obtains multiple user groups.Wherein, k-means clusters are calculated Clustering parameter k in method can be determined according to the mode of manual debugging.Principle based on clustering algorithm, is divided by clustering algorithm After user group, the user in same user group has larger region similitude, e.g., is all settled down in a line with the user in a group City, the user in different user group have less region similitude.
In the present embodiment, by the way that according to the corresponding relevant information of each user, to the whole network, user clusters, and obtains Multiple user groups, be able to will have lessly by the user clustering with larger region similitude into same user group The user clustering of domain similitude is into different user group.Due to the user behavior between the larger user of region similitude generally also More similar, the usual similitude of user behavior between the less user of region similitude is smaller, such as the one, user of tier 2 cities The program that purchases by group being commonly installed is rolled into a ball to be beautiful, and what the user in three line cities was commonly installed purchases by group program to spell the more, thus by for User divides user group, can be to realize that the accurate of application program recommends to provide basis.
It is mobile whole in above-mentioned steps S104 when the application program in the present embodiment recommends method by mobile terminal execution End obtains the application searches request of targeted customer, and judges whether the number of users of targeted customer place user group is more than predetermined number Amount.When the application program in the present embodiment recommends method to be performed by server, in above-mentioned steps S104, acquisition for mobile terminal mesh The application searches request of user is marked, and application searches request is sent to server, server and is getting the application searches During request, whether the number of users of user group is more than predetermined quantity where judging targeted customer.
In above-mentioned steps S106, if the number of users of user group is more than predetermined quantity where targeted customer, it is determined that target The application searches of user ask the category of employment of corresponding searched application.
In one embodiment, if application searches request includes the title of application program, by the title of the application program Corresponding application program is determined as searched application, mobile whole if application searches request includes the keyword of application program End or server determine include the application program of the keyword in title, and will be definite application program in installation is most answers With program, as searched application.
After definite application searches ask corresponding searched application, the acceptable description information according to searched application, Determine the category of employment of searched application, wherein, the category of employment of application program includes life kind, game class, audio-visual class etc..One In a specific embodiment, record has the category of employment of searched application in the description information of searched application.
In the present embodiment, predefining has targeted customer's to be currently located the similar row of corresponding first application program of country Table, which is used to record multiple set applications, and each set application is corresponding similar Using, also, the corresponding similar application of each set application recorded in the first application program affinity list, it is the current institute The application that user in country is installed.It should be noted that in the present embodiment, the user in each country, refers to growing Phase is settled down in the national user.
Correspondingly, in above-mentioned steps S108, the corresponding first application program phase of country is currently located according to targeted customer Recommend application program to targeted customer like list, be specially：
(a1) in above-mentioned first application program affinity list, it is similar to search searched application corresponding at least one first Using, also, according to the application program Generalization bounds based on collaborative filtering, determine that searched application is corresponding at least one Second similar application；
(a2) by order models trained in advance, application similar at least one first and at least one second phase seemingly should With being ranked up, target is recommended into the similar application of at least one first after sequence and at least one second similar application and used Family, wherein, the installation record training for the application program that order models recommend history based on targeted customer obtains.
In above-mentioned action (a1), in the first application program affinity list, search and be searched applying corresponding similar application, Using the similar application found as the first similar application, due to recorded in the first application program affinity list it is each it is set should Similar application, is the application program of the above-mentioned user installation being currently located in country, therefore the first similar application is also The application program of the above-mentioned user installation being currently located in country.
In above-mentioned action (a1), according to the application program Generalization bounds based on collaborative filtering, searched application is determined Corresponding at least one second similar application, alternatively, the collaborative filtering can be based on SVD (Singular value Decomposition, singular value decomposition) collaborative filtering.In the collaborative filtering based on SVD, by each user Rows of the ID (identy, identity) as two-dimensional matrix, the row of the ID of each application program as two-dimensional matrix, form one it is super Big two-dimensional matrix.If some application program of a certain user installation, by corresponding to the row of the user and the row of the application program Value be labeled as 1, be otherwise 0.To above-mentioned two-dimensional matrix carry out SVD decomposition after obtain two vectors, be respectively user vector and Application program vector, similar users and similar application can be determined by user's vector sum application program vector, so that it is determined that by Corresponding at least one second similar application is applied in search., can also be by User-based collaborative filterings in another embodiment Algorithm and Item-based collaborative filterings are combined, and replace the collaborative filtering based on SVD.
Since searched application is life kind application program, the first similar application and the second similar application are also life Class application program.
In above-mentioned action (a2), by the relevant information of at least one first similar application and at least one second similar application Relevant information input to order models trained in advance, and the application program ranking results of model output are obtained, according to this Application program ranking results, similar at least one first application and at least one second similar application are ranked up, and will row Targeted customer is recommended at least one first similar application and at least one second similar application after sequence.
Wherein, the installation record training for the application program that order models recommend history based on targeted customer obtains.Specifically Ground, obtains the installation record for the application program that targeted customer recommends history, is recorded according to the installation, determine that targeted customer is going through In the application program that history is recommended, the application program of installation, and uninstalled application program, and the application that installation will be selected are selected The relevant information of program is as positive sample, using the relevant information of uninstalled application program as negative sample, by the positive negative sample Input to Logic Regression Models are trained, and the model that training is completed is as above-mentioned order models.The above-mentioned application journey being related to The relevant information of sequence includes the description information of application program, can also include user's evaluation, user's marking etc..
In the present embodiment, the similar application of each set application recorded in the first application program affinity list is set, Application for the user installation being currently located in country, it is the above-mentioned use being currently located in country to enable to the first similar application The application of family installation, so that in the case where user searches for life kind application program, recommends the use in its country one belongs to user The similar application of family installation, so that user installation meets the life kind application program of its living environment, improves application program and recommends Precision.
In the present embodiment, with reference to the first application program affinity list and the application program based on collaborative filtering recommends plan Slightly, recommend application program to user, there is the beneficial effect for improving and recommending precision.
The installation record training for the application program recommended based on targeted customer history due to order models is obtained, and is led to Cross order models application similar to first and after the second similar application is ranked up, enable to the more forward application program that sorts More meet the behavioral characteristic of targeted customer's history installation application program, so as to preferentially recommend the application program for meeting user to user The application program of custom is installed, improves the application program search experience of user.
In the present embodiment, the corresponding second application program affinity list of place user group for having targeted customer is predefined, The second application program affinity list is used to record multiple set applications, and the corresponding phase of each set application seemingly should With, also, the corresponding similar application of each set application recorded in second the application list, it is place user The application being mounted in group, namely the application of the user installation in the place user group.
Correspondingly, it is similar according to corresponding second application program of the place user group of targeted customer in above-mentioned steps S108 List recommends application program to targeted customer, is specially：
(b1) in above-mentioned second application program affinity list, the corresponding at least one third phase of searched application is searched seemingly Using, also, according to the application program Generalization bounds based on collaborative filtering, determine that searched application is corresponding at least one 4th similar application；
(b2) by order models trained in advance, at least one third phase seemingly should be used and at least one 4th phase seemingly should With being ranked up, at least one third phase after sequence seemingly should be used with target is recommended with least one 4th similar application Family, wherein, the installation record training for the application program that order models recommend history based on targeted customer obtains.
In above-mentioned action (b1), in the second application program affinity list, search and be searched applying corresponding similar application, The similar application found seemingly should be used as third phase, due to recorded in the second application program affinity list it is each it is set should Similar application, is the application of the user installation in above-mentioned place user group, therefore it is also above-mentioned institute that third phase, which seemingly should be used, The application of user installation in user group.
In above-mentioned action (b1), according to the above-mentioned application program Generalization bounds based on collaborative filtering, determine searched Using corresponding at least one 4th similar application.The process may be referred to the description of foregoing (a1), not do repetition here.
Since searched application is non-life kind application program, third phase seemingly should be used and the 4th similar application is also to be non- Life kind application program.
In above-mentioned action (b2), by least one third phase seemingly should relevant information and at least one 4th similar application Relevant information input to order models trained in advance, and the application program ranking results of model output are obtained, according to this Application program ranking results, seemingly should be ranked up at least one third phase with least one 4th similar application, and will row At least one third phase after sequence seemingly should recommend targeted customer with least one 4th similar application.Can on order models With with reference to description above, which is not described herein again.
In the present embodiment, the similar application of each set application recorded in the second application program affinity list is set, For the application of the user installation in the place user group, third phase is enabled to use as the user in above-mentioned place user group The application of installation, so that in the case where user searches for non-life kind application program, recommends to meet its user classification feelings to user The application program of condition, so that user installation meets the application program of its user classification situation, improves the accurate of application program recommendation Degree.
In the present embodiment, with reference to the second application program affinity list and the application program based on collaborative filtering recommends plan Slightly, recommend application program to user, there is the beneficial effect for improving and recommending precision.
The installation record training for the application program recommended based on targeted customer history due to order models is obtained, and is led to Cross order models to third phase seemingly should with and after the 4th similar application is ranked up, enable to the more forward application program that sorts More meet the behavioral characteristic of targeted customer's history installation application program, so as to preferentially recommend the application program for meeting user to user The application program of custom is installed, improves the application program search experience of user.
Since life kind application program (such as cuisines take-away program) needs to match with current living environment, in order to user Using life kind application program, non-life kind application program (such as read routine) need not match with current living environment, need To match with the behavioural habits of user itself, to improve the usage experience of user, therefore in the present embodiment, life kind will be searched for Separately treated using with the application of non-life kind, divide situation to recommend application program.In search life kind in application, being based on targeted customer The corresponding first application program affinity list of country that is currently located recommend into line program, enable to user installation to live with it The life kind application program of environments match, improves the precision that application program is recommended；Non- life kind is being searched in application, being based on mesh The corresponding second application program affinity list of place user group for marking user is recommended into line program, enables to user installation and its The matched application program of user behavior, improves the precision that application program is recommended.
In the present embodiment, before step S108, it is also necessary to determine that targeted customer's is currently located country corresponding first Application program affinity list, specific determination process are：
(c1) predetermined application program affinity list is obtained；Wherein, in the application program affinity list there is full record Each application in net, and each apply corresponding similar application；Wherein, each application in the whole network is referred in internet Whole applications；
(c2) the application program affinity list record each apply corresponding similar application in, deletion be currently located state The uninstalled application program of user in family；
(c3) by the application program affinity list after deletion, as being currently located, corresponding first application program of country is similar List.
Specifically, predetermined application program affinity list is obtained, alternatively, which includes mesh Whole application programs existing for preceding the whole network, to ensure the comprehensive of application program, which has also recorded entirely The corresponding similar application of each application program existing for net, the partial content of the application program affinity list can be such as table 1 below institute Show.
Table 1
Application program Similar application
Wechat Microblogging, facebook
Iqiyi.com Youku.com, youtube
Act in (c2), in the corresponding similar application of each application program of application program affinity list record, delete Except the uninstalled application program of the user being currently located in country.Specifically, each national application program can be counted in advance List is installed, application program installation list is used for the popular application program for recording the user installation in corresponding country, will be current The application program installation list of the country one belongs to is compared with above application program affinity list, remembers in application program affinity list In the corresponding similar application of each application program of record, delete Unrecorded in the application program installation list for be currently located country Application program, so as to delete the uninstalled application program of the user being currently located in country.
The application program installation list of each country can use to be obtained in the following ways：To Mr. Yu country, it is every to count the state The application program that a user is installed, so that it is determined that the application program that the user in the country is installed, determines installation highest N application program, n can be 50,000 or other numerical value, and the highest n application program of installation is dropped according to installation Sequence sorts, and obtains the national application program installation list.
Act in (c3), the application program affinity list after deletion is applied as country corresponding first is currently located Program affinity list.Due to deleting the uninstalled application program of user being currently located in country by acting (c2), The similar application recorded in first application program affinity list, is the application program for the user installation being currently located in country.
In the present embodiment, the corresponding first application program affinity list of country is currently located by obtaining, can be easy to User recommends the application program of the user installation in its country one belongs to, so as to avoid application program and the user recommended to user The situation that living environment is not inconsistent.Such as if it is the U.S. that certain user, which is currently located country, when it searches for wechat, for its recommendation Facebook, if certain user is currently located country as China, when it searches for wechat, recommends microblogging for it.
In the present embodiment, each national application program installation list can be counted in advance, and answer according to each national List and above-mentioned application program affinity list are installed with program, determine the similar row of corresponding first application program of each country Table, consequently facilitating quickly recommending application program to user.
Correspondingly, in the present embodiment, before step S108, it is also necessary to determine that the place user group of targeted customer is corresponding Second application program affinity list, specific determination process are：
(d1) predetermined application program affinity list is obtained；Wherein, record has the whole network in application program affinity list In each application, and each apply corresponding similar application；Application in the application program affinity list, that is, above-mentioned (c1) Program affinity list；
(d2) corresponding similar application each is applied in application program affinity list record, where deleting in user group The uninstalled application program of user；
(d3) by the application program affinity list after deletion, as the similar row of corresponding second application program of place user group Table.
Specifically, as shown in above-mentioned (c1), predetermined application program affinity list is obtained, alternatively, this applies journey Sequence affinity list includes whole application programs existing for current the whole network, to ensure the comprehensive of application program, the application program phase The corresponding similar application of each application program existing for also having recorded the whole network like list, in the part of the application program affinity list Appearance can be as listed in Table 1.
Act in (d2), in the corresponding similar application of each application program of application program affinity list record, delete Except the uninstalled application program of user in the user group of place.Specifically, the application program of each user group can be counted in advance List is installed, application program installation list is used for the popular application program for recording the user installation in corresponding user group, by institute It is compared with above application program affinity list in the application program installation list of user group, remembers in application program affinity list It is Unrecorded in the application program installation list of user group where deleting to answer in the corresponding similar application of each application program of record With program, so that the uninstalled application program of user in user group where deleting.
The application program installation list of each user group can use to be obtained in the following ways：To Mr. Yu's user group, statistics should The application program that each user of user group is installed, so that it is determined that the application program that the user in the user group is installed, determines The highest m application program of installation, m can be 50,000 or other numerical value, by the highest m application program of installation according to peace Loading amount carries out descending sort, obtains the application program installation list of the user group.
Act in (d3), by the application program affinity list after deletion, journey is applied as place user group corresponding second Sequence affinity list.The uninstalled application program of user in the user group where being deleted by action (d2), second The similar application recorded in application program affinity list, is the application program of the user installation in the user group of place.
In the present embodiment, by the corresponding second application program affinity list of user group where obtaining, can be easy to The application program of the user installation in user group where it is recommended at family, so as to avoid the application program recommended to user and user The situation that classification situation is not inconsistent.For example for body-building application, the user group where targeted customer uses keep application journeys mostly Sequence, therefore body-building is searched in application, recommending keep application programs to targeted customer in targeted customer.
In the present embodiment, the application program installation list of each user group can be counted in advance, and according to each user group Application program installation list and above-mentioned application program affinity list, determine the corresponding second application program phase of each user group Like list, consequently facilitating quickly recommending application program to user.
In the present embodiment, to user recommend application program when, except the searching request according to user to user recommend apply Beyond program, the application program that can also be installed in the recent period according to user, recommends application program to user.For example detection user is most The application program of installation in nearly one week, when detecting life kind application program, in the first application program affinity list, searches inspection The similar application of the corresponding at least one first kind of the application program that measures, recommends according to the application program based on collaborative filtering Strategy, determines the similar application of corresponding at least one second class of the application program that detects, by order models trained in advance, Similar with least one second class application of application similar at least one first kind is ranked up, by least one the after sequence User is recommended in the similar application of one kind application similar with least one second class.Similarly, non-life kind application program is being detected When, in the same fashion, pushed away according to above-mentioned second application program affinity list and the application program based on collaborative filtering Strategy is recommended, recommends application program to user, so as to increase the diversity that application program recommends scene so that user opens every time should During with download program software (such as software store or Google Play), recommended application program may browse through.
In the embodiment of the present application, by the presentation of information of recommended application program on the interface of the mobile terminal of user, So as to recommend application program to user.
By the generating process of above-mentioned first application program affinity list and the second application program affinity list, first The generation of application program affinity list and the second application program affinity list all relies on foregoing application program affinity list, because , to ensure that application program affinity list includes more application program, preferred application affinity list is deposited including current the whole network for this Whole application programs, to ensure the comprehensive of application program.In above-mentioned generation the first application program affinity list and second Before application program affinity list, the application program affinity list can also be generated in the following manner：
(e1) whole application programs of the whole network user installation are determined, and apply the mark in shop to believe according to GooglePlay Breath, captures the corresponding 5th similar list of application of each application program in GooglePlay applies shop, generates the 5th list, In 5th list, according to the installation order from high to low of each application program, each application program is ranked up, and remembers Record the similar application of each application program；
(e2) in GooglePlay applies shop, determine that GooglePlay applies retouching for each application program in shop Information is stated, extraction of semantics is carried out to the description information of each application program using textrank, tfidf, rake scheduling algorithm, and And NER name Entity recognitions are carried out, the name entity such as name of product in description information is labeled, by extraction of semantics knot Fruit and annotation results are combined, and obtain the label of each application program, recycle word embedding word embedded mobile GISs The label of each application program is subjected to vectorization expression, is represented according to the vectorization of the label of each application program as a result, really Semantic similarity between the label of fixed each application program, according to the semantic similarity between the label of each application program, Determine the similar application of each application program, generate the 6th list, in the 6th list, according to each application program installation by High to low order, is ranked up each application program, and records the similar application of each application program；
(e3) above-mentioned 5th list and above-mentioned 6th list are combined, the program that is applied affinity list.Specifically, The highest p application program of installation in above-mentioned 5th list (or all applications) is taken, and the phase of each application program seemingly should With, same the, the highest p application program of installation in above-mentioned 6th list (or all applications) is taken, and each apply journey The similar application of sequence, the application program translocation sorting that will be taken out in the application program taken out in the 5th list and the 6th list, obtains To application program affinity list, such as, A, B, C are taken out in the 5th list, first, second, third are taken out in the 6th list, then application program Affinity list is A, first, B, second, C, third.Each application program is arranged on earth according to installation by height in application program affinity list Sequence, and the similar application of each application program is recorded, p can take 50,000 or other numerical value, be taken from the 5th or the 6th list The form of the data gone out is specifically as follows A1-A2, A3, wherein, A1 is the mark of application program, and A3, A2 are the similar application of A1.
In the present embodiment, with reference to the whole network user installation whole application programs and GooglePlay apply shop in whole Application program, generates application program affinity list, can ensure that application program affinity list is included existing for current the whole network all Application program, to ensure the comprehensive of application program.
In the present embodiment, before step S102, it can determine user's according to the location information of the mobile terminal of user The country one belongs to, and, determine that the colonization of user is national in the following manner：
(1) application program for the being currently located country installation list of user is obtained；Wherein, remember in application program installation list Record has the multiple applications for the user installation being currently located in country；
(2) according to the matching degree between application program installation list, and the application program installation list of user, judge to use Family be currently located country whether be user colonization country；The application program installation list of user is installed complete including user Portion's application program；
(3) if so, user then is currently located the national colonization country for being determined as user, otherwise, according to other countries Application program installation list, and the matching degree between the installation list of the application program of user, determines the colonization country of user.
As outlined above, in the present embodiment, each national application program installation list can be counted in advance, this applies journey Sequence installation list is used for the popular application program for recording the user installation in corresponding country.In the present embodiment, working as user is obtained The application program installation list of preceding the country one belongs to, and determine application program installation list, and the application program installation row of user Matching degree between table, the matching degree refer to that application program installation list is overlapped with the application program installation list of user Application program quantity, take family application program installation list in application program quantity ratio, if the matching degree is big In preset value, it is determined that user's is currently located colonization country of the country for user, and the country that is currently located of user is determined as The colonization country of user, otherwise, calculates the application program installation list of other countries, and the application program installation row of user respectively Matching degree between table, by the corresponding country of application program installation list of matching degree maximum, the colonization for being determined as user is national. Using this kind of mode, it may be determined that the colonization country of targeted customer.
In the above process, matching degree calculated example is the application program installation list of application program installation list and user The quantity of the application program of middle coincidence is 20, and the quantity of application program is 30 in the application program installation list of user, then Matching degree is 20/30=66.7%.
Due to settling down the user in certain state, the application program that its application program used should be used with the user in the state It is more consistent, therefore in the present embodiment, according to the application program of user install list and country application program installation list it Preceding matching degree, determines the colonization country of user, has the advantages that definite effect is accurate.By determine user colonization country and The country one belongs to of user, can distinguish the situation that user lives in certain state in short term, so as to carry out accurately user group division.
, will be complete when the number of users of the place user group of targeted customer is less than or equal to predetermined quantity in the present embodiment The corresponding application program installation list jointly of portion's user group, recommends targeted customer；Wherein, whole user groups are jointly corresponding should Installed with program in list, according to the installation of application program, descending record there are the multiple of the user installation in whole user groups Application program.
In the present embodiment, for the user in whole user groups, the application program installed according to each user, determines complete The application program that portion is mounted, and determine each application program installation, determine installation highest 50,000 (it is of course possible to for Other predetermined values) a application program, and this certain applications program is arranged according to installation descending, obtain whole user groups and be total to List is installed with corresponding application program, is less than or equal to predetermined quantity in the number of users of the place user group of targeted customer When (such as 10,000), targeted customer is recommended into application program installation list, so that the program based on the user in whole user groups Installation situation, recommends application program to targeted customer.
Fig. 2 is the flow diagram that the application program that another embodiment of the application provides recommends method, as shown in Fig. 2, should Method comprises the following steps：
Step S202, obtains the application searches request of user.
Step S204, judges whether the number of users of the user group where user is more than default quantity.
If more than, then step S206 is performed, otherwise, execution step S212.
Step S206, judges that the application searches ask whether the classification of corresponding searched application is life kind.
If so, performing step S208, otherwise, step S210 is performed.
Step S208, according to the national corresponding first application program affinity list that is currently located of user, and, based on association With the application program Generalization bounds of filter algorithm, recommend application program to user.
Step S210, according to the corresponding second application program affinity list of the place user group of user, and, based on collaboration The application program Generalization bounds of filter algorithm, recommend application program to user.
Step S212, by whole user groups, corresponding application program installs list jointly, recommends user.
To sum up, the embodiment of the present application has the following advantages：
(1) according to a variety of relevant informations of user, colony's division is carried out to user, improves the accurate of application program recommendation Property, avoid three or four line cities recommend whats app, religion area recommend live application situations such as occur；
(2) user is currently located state and adoptive country as one of factor of user clustering, can be to living in short term The user of country more accurately recommends application program beyond adoptive country；
(3) life kind is applied and non-life kind application division is treated, the application program for improving user recommends experience.
Corresponding above-mentioned method, the embodiment of the present application additionally provide a kind of application program recommendation apparatus, and Fig. 3 is the application one The module composition schematic diagram for the application program recommendation apparatus that embodiment provides, as shown in figure 3, the device includes：
User clustering module 41, for according to the corresponding relevant information of each user, the whole network user being divided into more A user group；Wherein, the relevant information is national, fixed including religious belief, place city, the country one belongs to, language used, colonization Occupy at least one in city, the income level in the place city, the income level for settling down city；
Quantity discrimination module 42, for when receiving the application searches request of targeted customer, judging the targeted customer Whether the number of users of place user group is more than predetermined quantity；
Classification discrimination module 43, for if more than predetermined quantity, it is determined that the application searches request pair of the targeted customer The category of employment for the searched application answered；
Be life kind application for such as described category of employment using recommending module 44, then according to the targeted customer work as The corresponding first application program affinity list of preceding the country one belongs to recommends application program, such as category of employment to the targeted customer For non-life kind application, then according to the corresponding second application program affinity list of place user group of the targeted customer to described Targeted customer recommends application program.
, being capable of the searching request based on user, the place user group of user and the place of user by the embodiment of the present application Country, personalizedly recommends application program to user, improves the recommendation precision of application program.Also, search for and live in user Class to user based on the country one belongs to of user in application, recommend application program, the application program for enabling to recommend meets user The country one belongs to habits and customs so that improve user application searches experience.Also, search for non-life kind application in user When, the place user group based on user recommends application program to user, enables to the place of the application program and user recommended User group matches, i.e., the classification situation with user matches, so as to improve the application searches experience of user.
Alternatively, the user clustering module 41 is specifically used for：
According to default clustering algorithm, according to the corresponding relevant information of each user, to the whole network user Clustered, obtain multiple user groups.
Alternatively, the first application program affinity list is used to record multiple set applications, and each set application Corresponding similar application；Wherein, the corresponding similar application of the set application is described to be currently located user in country The application installed；
The application recommending module 44 is specifically used for：
In the first application program affinity list, it is corresponding at least one first similar to search the searched application Using, also, Generalization bounds are applied according to based on collaborative filtering, determine that the searched application is corresponding at least one Second similar application；
By order models trained in advance, at least one first similar application and at least one second phase At least one first similar application after sequence and at least one second similar application seemingly should be pushed away with being ranked up Recommend to the targeted customer；Wherein, the installation for the application program that the order models recommend history based on the targeted customer Record training obtains.
Alternatively, the second application program affinity list is used to record multiple set applications, and each set application Corresponding similar application；Wherein, the corresponding similar application of the set application is to be mounted in the place user group Application；
The application recommending module 44 is specifically used for：
In the second application program affinity list, the corresponding at least one third phase of the searched application is searched seemingly Using, also, Generalization bounds are applied according to based on collaborative filtering, determine that the searched application is corresponding at least one 4th similar application；
By order models trained in advance, at least one third phase seemingly should be used and at least one 4th phase At least one third phase after sequence seemingly should seemingly should be pushed away with least one 4th similar application with being ranked up Recommend to the targeted customer；Wherein, the installation for the application program that the order models recommend history based on the targeted customer Record training obtains.
Fig. 4 is the module composition schematic diagram for the application program recommendation apparatus that another embodiment of the application provides, such as Fig. 4 institutes Show, alternatively, further include first list determining module 51, be used for：
Obtain predetermined application program affinity list；Wherein, record has the whole network in the application program affinity list In each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, be currently located described in deletion The uninstalled application program of user in country；
By the application program affinity list after deletion, corresponding first application program of country is currently located as described Affinity list.
As shown in figure 4, alternatively, second list determining module 52 is further included, is used for：
Obtain predetermined application program affinity list；Wherein, record has the whole network in the application program affinity list In each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, delete the place user The uninstalled application program of user in group；
By the application program affinity list after deletion, as the corresponding second application program phase of the place user group Like list.
Alternatively, adoptive country's determining module is further included, is used for：
Obtain the application program for the being currently located country installation list of user；Wherein, in the application program installation list Record has multiple applications of the user installation being currently located in country；
Matching degree between list, and the application program installation list of the user is installed according to the application program, is sentenced Break the user be currently located country whether be the user colonization country；
If so, the user is then currently located the national colonization country for being determined as the user, otherwise, according to other The application program installation list of country, the matching degree between the application program installation list of the user, determines the user Colonization country.
Further, based on above-mentioned method, the embodiment of the present application additionally provides a kind of application program recommendation apparatus, Fig. 5 The structure diagram of the application program recommendation apparatus provided for one embodiment of the application.
As shown in figure 5, application program recommendation apparatus can produce bigger difference because configuration or performance are different, can wrap One or more processor 701 and memory 702 are included, can be stored with one or more in memory 702 deposits Store up application program or data.Wherein, memory 702 can be of short duration storage or persistently storage.It is stored in the application of memory 702 Program can include one or more modules (diagram is not shown), and each module can include to application program recommendation apparatus In series of computation machine executable instruction.Further, processor 701 could be provided as communicating with memory 702, answer With the series of computation machine executable instruction performed on program recommendation apparatus in memory 702.Application program recommendation apparatus may be used also With including one or more power supplys 703, one or more wired or wireless network interfaces 704, one or one with Upper input/output interface 705, one or more keyboards 706 etc..
In a specific embodiment, application program recommendation apparatus includes memory, processor and is stored in described deposit On reservoir and the computer program that can run on the processor, the computer program can when being performed by the processor Realize the process as described in above-mentioned application program recommends embodiment of the method, specifically include following steps：
According to the corresponding relevant information of each user, the whole network user is divided into multiple user groups；Wherein, the phase Closing information includes religious belief, place city, the country one belongs to, language used, colonization country, colonization city, the place city Income level, it is described settle down city income level at least one of；
When the application searches for receiving targeted customer are asked, the number of users of user group where judging the targeted customer Whether predetermined quantity is more than；
If more than predetermined quantity, it is determined that the application searches of the targeted customer ask the industry of corresponding searched application Classification；
Such as category of employment is life kind application, then is currently located country corresponding first according to the targeted customer Application program affinity list recommends application program to the targeted customer, and such as category of employment is non-life kind application, then root According to the corresponding second application program affinity list of the place user group of the targeted customer journey is applied to targeted customer recommendation Sequence.
Alternatively, when the computer program is performed by the processor, according to the corresponding related letter of each user Breath, multiple user groups are divided into by the whole network user, including：
According to default clustering algorithm, according to the corresponding relevant information of each user, to the whole network user Clustered, obtain multiple user groups.
Alternatively, when the computer program is performed by the processor, the first application program affinity list is used for Record multiple set applications, and the corresponding similar application of each set application；Wherein, the corresponding phase of the set application The application that user in country is installed seemingly should be currently located described in being；
Used according to the corresponding first application program affinity list of country that is currently located of the targeted customer to the target Application program is recommended at family, including：
In the first application program affinity list, it is corresponding at least one first similar to search the searched application Using, also, Generalization bounds are applied according to based on collaborative filtering, determine that the searched application is corresponding at least one Second similar application；
By order models trained in advance, at least one first similar application and at least one second phase At least one first similar application after sequence and at least one second similar application seemingly should be pushed away with being ranked up Recommend to the targeted customer；Wherein, the installation for the application program that the order models recommend history based on the targeted customer Record training obtains.
Alternatively, when the computer program is performed by the processor, the second application program affinity list is used for Record multiple set applications, and the corresponding similar application of each set application；Wherein, the corresponding phase of the set application Seemingly should be in the place user group be mounted application；
According to the corresponding second application program affinity list of the place user group of the targeted customer to the targeted customer Recommend application program, including：
In the second application program affinity list, the corresponding at least one third phase of the searched application is searched seemingly Using, also, Generalization bounds are applied according to based on collaborative filtering, determine that the searched application is corresponding at least one 4th similar application；
By order models trained in advance, at least one third phase seemingly should be used and at least one 4th phase At least one third phase after sequence seemingly should seemingly should be pushed away with least one 4th similar application with being ranked up Recommend to the targeted customer；Wherein, the installation for the application program that the order models recommend history based on the targeted customer Record training obtains.
Alternatively, when the computer program is performed by the processor, further include：
Obtain predetermined application program affinity list；Wherein, record has the whole network in the application program affinity list In each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, be currently located described in deletion The uninstalled application program of user in country；
By the application program affinity list after deletion, corresponding first application program of country is currently located as described Affinity list.
Alternatively, when the computer program is performed by the processor, further include：
Obtain predetermined application program affinity list；Wherein, record has the whole network in the application program affinity list In each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, delete the place user The uninstalled application program of user in group；
By the application program affinity list after deletion, as the corresponding second application program phase of the place user group Like list.
Alternatively, when the computer program is performed by the processor, further include：
Obtain the application program for the being currently located country installation list of user；Wherein, in the application program installation list Record has multiple applications of the user installation being currently located in country；
Matching degree between list, and the application program installation list of the user is installed according to the application program, is sentenced Break the user be currently located country whether be the user colonization country；
If so, the user is then currently located the national colonization country for being determined as the user, otherwise, according to other The application program installation list of country, the matching degree between the application program installation list of the user, determines the user Colonization country.
, being capable of the searching request based on user, the place user group of user and the place of user by the embodiment of the present application Country, personalizedly recommends application program to user, improves the recommendation precision of application program.Also, search for and live in user Class to user based on the country one belongs to of user in application, recommend application program, the application program for enabling to recommend meets user The country one belongs to habits and customs so that improve user application searches experience.Also, search for non-life kind application in user When, the place user group based on user recommends application program to user, enables to the place of the application program and user recommended User group matches, i.e., the classification situation with user matches, so as to improve the application searches experience of user.
Further, the embodiment of the present application additionally provides a kind of storage medium, for storing computer executable instructions, one In kind specific embodiment, which can be USB flash disk, CD, hard disk etc., and the computer of storage medium storage can perform Instruction can realize the process as described in above-mentioned application program recommends embodiment of the method, and reach phase when being executed by processor Same effect, which is not described herein again.
Each embodiment in this specification is described by the way of progressive, identical similar portion between each embodiment Divide mutually referring to what each embodiment stressed is the difference with other embodiment.It is real especially for system For applying example, since it is substantially similar to embodiment of the method, so description is fairly simple, related part is referring to embodiment of the method Part explanation.
The foregoing is merely embodiments herein, is not limited to the application.For those skilled in the art For, the application can have various modifications and variations.All any modifications made within spirit herein and principle, be equal Replace, improve etc., it should be included within the scope of claims hereof.

Claims (14)

1. a kind of application program recommends method, it is characterised in that including：
According to the corresponding relevant information of each user, the whole network user is divided into multiple user groups；Wherein, the related letter Breath includes religious belief, place city, the country one belongs to, language used, colonization country, colonization city, the receipts in the place city Enter at least one in horizontal, the described income level for settling down city；
When the application searches for receiving targeted customer are asked, whether the number of users of user group where judging the targeted customer More than predetermined quantity；
If more than predetermined quantity, it is determined that the application searches of the targeted customer ask the industry class of corresponding searched application Not；
Such as category of employment is life kind application, then is currently located corresponding first application of country according to the targeted customer Program affinity list recommends application program to the targeted customer, and such as category of employment is non-life kind application, then according to institute The corresponding second application program affinity list of place user group for stating targeted customer recommends application program to the targeted customer.
2. the method as described in claim 1, it is characterised in that according to the corresponding relevant information of each user, by the whole network User is divided into multiple user groups, including：
According to default clustering algorithm, according to the corresponding relevant information of each user, the whole network user is carried out Cluster, obtains multiple user groups.
3. the method as described in claim 1, it is characterised in that the first application program affinity list be used for record it is multiple both Fixed application, and the corresponding similar application of each set application；Wherein, the corresponding similar application of the set application is The application for being currently located user in country and being installed；
Pushed away according to the corresponding first application program affinity list of country that is currently located of the targeted customer to the targeted customer Application program is recommended, including：
In the first application program affinity list, searching corresponding at least one first phase of the searched application seemingly should With, also, Generalization bounds are applied according to based on collaborative filtering, determine the searched application corresponding at least one the Two-phase seemingly should be used；
, seemingly should at least one first similar application and at least one second phase by order models trained in advance With being ranked up, at least one first similar application after sequence and at least one second similar application are recommended The targeted customer；Wherein, the installation record for the application program that the order models recommend history based on the targeted customer Training obtains.
4. the method as described in claim 1, it is characterised in that the second application program affinity list be used for record it is multiple both Fixed application, and the corresponding similar application of each set application；Wherein, the corresponding similar application of the set application is The application being mounted in the place user group；
Recommended according to the corresponding second application program affinity list of the place user group of the targeted customer to the targeted customer Application program, including：
In the second application program affinity list, searching the corresponding at least one third phase of the searched application seemingly should With, also, Generalization bounds are applied according to based on collaborative filtering, determine the searched application corresponding at least one the Four similar applications；
By order models trained in advance, at least one third phase seemingly should be used and at least one 4th phase seemingly should With being ranked up, at least one third phase after sequence seemingly should be recommended with least one 4th similar application The targeted customer；Wherein, the installation record for the application program that the order models recommend history based on the targeted customer Training obtains.
5. the method as described in claim 1, it is characterised in that further include：
Obtain predetermined application program affinity list；Wherein, being recorded in the application program affinity list has in the whole network Each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, country is currently located described in deletion In the uninstalled application program of user；
By the application program affinity list after deletion, to be currently located corresponding first application program of country similar as described List.
6. the method as described in claim 1, it is characterised in that further include：
Obtain predetermined application program affinity list；Wherein, being recorded in the application program affinity list has in the whole network Each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, delete in the place user group The uninstalled application program of user；
By the application program affinity list after deletion, as the similar row of corresponding second application program of the place user group Table.
7. such as claim 1 to 6 any one of them method, it is characterised in that further include：
Obtain the application program for the being currently located country installation list of user；Wherein, recorded in the application program installation list There are multiple applications of the user installation being currently located in country；
Matching degree between list, and the application program installation list of the user is installed according to the application program, judges institute State user be currently located country whether be the user colonization country；
If so, the user is then currently located the national colonization country for being determined as the user, otherwise, according to other countries Application program installation list, the matching degree between the installation list of the application program of the user, determines determining for the user Occupy country.
A kind of 8. application program recommendation apparatus, it is characterised in that including：
User clustering module, for according to the corresponding relevant information of each user, the whole network user to be divided into multiple users Group；Wherein, the relevant information includes religious belief, place city, the country one belongs to, language used, colonization country, colonization city At least one of in city, the income level in the place city, the income level for settling down city；
Quantity discrimination module, for when receiving the application searches request of targeted customer, judging to use where the targeted customer Whether the number of users of family group is more than predetermined quantity；
Classification discrimination module, for if more than predetermined quantity, it is determined that the application searches of the targeted customer ask corresponding quilt Search for the category of employment of application；
Be life kind application for such as described category of employment using recommending module, then being currently located according to the targeted customer The corresponding first application program affinity list of country recommends application program to the targeted customer, and such as category of employment is non-life Class application living, then use according to the corresponding second application program affinity list of place user group of the targeted customer to the target Recommend application program in family.
9. device as claimed in claim 8, it is characterised in that the user clustering module is specifically used for：
According to default clustering algorithm, according to the corresponding relevant information of each user, the whole network user is carried out Cluster, obtains multiple user groups.
10. device as claimed in claim 8, it is characterised in that the first application program affinity list is used to record multiple Set application, and the corresponding similar application of each set application；Wherein, the corresponding similar application of the set application is equal By the application for being currently located user in country and installing；
The application recommending module is specifically used for：
In the first application program affinity list, searching corresponding at least one first phase of the searched application seemingly should With, also, Generalization bounds are applied according to based on collaborative filtering, determine the searched application corresponding at least one the Two-phase seemingly should be used；
, seemingly should at least one first similar application and at least one second phase by order models trained in advance With being ranked up, at least one first similar application after sequence and at least one second similar application are recommended The targeted customer；Wherein, the installation record for the application program that the order models recommend history based on the targeted customer Training obtains.
11. device as claimed in claim 8, it is characterised in that the second application program affinity list is used to record multiple Set application, and the corresponding similar application of each set application；Wherein, the corresponding similar application of the set application is equal For the application being mounted in the place user group；
The application recommending module is specifically used for：
In the second application program affinity list, searching the corresponding at least one third phase of the searched application seemingly should With, also, Generalization bounds are applied according to based on collaborative filtering, determine the searched application corresponding at least one the Four similar applications；
By order models trained in advance, at least one third phase seemingly should be used and at least one 4th phase seemingly should With being ranked up, at least one third phase after sequence seemingly should be recommended with least one 4th similar application The targeted customer；Wherein, the installation record for the application program that the order models recommend history based on the targeted customer Training obtains.
12. device as claimed in claim 8, it is characterised in that further include first list determining module, be used for：
Obtain predetermined application program affinity list；Wherein, being recorded in the application program affinity list has in the whole network Each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, country is currently located described in deletion In the uninstalled application program of user；
By the application program affinity list after deletion, to be currently located corresponding first application program of country similar as described List.
13. device as claimed in claim 8, it is characterised in that further include second list determining module, be used for：
Obtain predetermined application program affinity list；Wherein, being recorded in the application program affinity list has in the whole network Each application, and each apply corresponding similar application；
The application program affinity list record each apply corresponding similar application in, delete in the place user group The uninstalled application program of user；
By the application program affinity list after deletion, as the similar row of corresponding second application program of the place user group Table.
14. such as claim 8 to 13 any one of them device, it is characterised in that further include adoptive country's determining module, be used for：
Obtain the application program for the being currently located country installation list of user；Wherein, recorded in the application program installation list There are multiple applications of the user installation being currently located in country；
Matching degree between list, and the application program installation list of the user is installed according to the application program, judges institute State user be currently located country whether be the user colonization country；
If so, the user is then currently located the national colonization country for being determined as the user, otherwise, according to other countries Application program installation list, the matching degree between the installation list of the application program of the user, determines determining for the user Occupy country.
CN201711307219.6A 2017-12-11 2017-12-11 Application program recommendation method and device Active CN107977445B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711307219.6A CN107977445B (en) 2017-12-11 2017-12-11 Application program recommendation method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711307219.6A CN107977445B (en) 2017-12-11 2017-12-11 Application program recommendation method and device

Publications (2)

Publication Number Publication Date
CN107977445A true CN107977445A (en) 2018-05-01
CN107977445B CN107977445B (en) 2022-05-03

Family

ID=62009919

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711307219.6A Active CN107977445B (en) 2017-12-11 2017-12-11 Application program recommendation method and device

Country Status (1)

Country Link
CN (1) CN107977445B (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108830298A (en) * 2018-05-21 2018-11-16 麒麟合盛网络技术股份有限公司 A kind of method and device of determining user characteristics label
CN109255606A (en) * 2018-08-31 2019-01-22 阿里巴巴集团控股有限公司 Default recommended method, device and the payment terminal of the means of payment
CN109408679A (en) * 2018-09-28 2019-03-01 平安科技（深圳）有限公司 Method, apparatus, electronic equipment and the storage medium of intelligent management application program
CN109582865A (en) * 2018-11-19 2019-04-05 北京奇虎科技有限公司 A kind of method and device of pushing application program
CN110175290A (en) * 2019-05-24 2019-08-27 竞技世界(成都)网络技术有限公司 A kind of content delivery method and device
CN111859149A (en) * 2020-08-03 2020-10-30 腾讯科技（北京）有限公司 Information recommendation method and device, electronic equipment and storage medium
CN112650940A (en) * 2019-10-10 2021-04-13 北京多点在线科技有限公司 Recommendation method and device of application program, computer equipment and storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103617543A (en) * 2013-11-25 2014-03-05 百度在线网络技术（北京）有限公司 Recommendation method and device of application programs
US20170109151A1 (en) * 2015-10-16 2017-04-20 International Business Machines Corporation System and method for context aware mobile application installation queuing
CN107273544A (en) * 2017-07-07 2017-10-20 广东小天才科技有限公司 Application program recommendation method, application program recommendation device and server

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103617543A (en) * 2013-11-25 2014-03-05 百度在线网络技术（北京）有限公司 Recommendation method and device of application programs
US20170109151A1 (en) * 2015-10-16 2017-04-20 International Business Machines Corporation System and method for context aware mobile application installation queuing
CN107273544A (en) * 2017-07-07 2017-10-20 广东小天才科技有限公司 Application program recommendation method, application program recommendation device and server

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108830298A (en) * 2018-05-21 2018-11-16 麒麟合盛网络技术股份有限公司 A kind of method and device of determining user characteristics label
CN108830298B (en) * 2018-05-21 2021-07-23 麒麟合盛网络技术股份有限公司 Method and device for determining user feature tag
CN109255606A (en) * 2018-08-31 2019-01-22 阿里巴巴集团控股有限公司 Default recommended method, device and the payment terminal of the means of payment
CN109408679A (en) * 2018-09-28 2019-03-01 平安科技（深圳）有限公司 Method, apparatus, electronic equipment and the storage medium of intelligent management application program
CN109582865A (en) * 2018-11-19 2019-04-05 北京奇虎科技有限公司 A kind of method and device of pushing application program
CN110175290A (en) * 2019-05-24 2019-08-27 竞技世界(成都)网络技术有限公司 A kind of content delivery method and device
CN110175290B (en) * 2019-05-24 2022-03-15 竞技世界(成都)网络技术有限公司 Content pushing method and device
CN112650940A (en) * 2019-10-10 2021-04-13 北京多点在线科技有限公司 Recommendation method and device of application program, computer equipment and storage medium
CN111859149A (en) * 2020-08-03 2020-10-30 腾讯科技（北京）有限公司 Information recommendation method and device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN107977445B (en) 2022-05-03

Similar Documents

Publication Publication Date Title
CN107977445A (en) Application program recommends method and device
CN102708130B (en) Calculate the easily extensible engine that fine point of user is mated for offer
CN110574387B (en) Recommending live streaming content using machine learning
CN106874279B (en) Method and device for generating application category label
US10180967B2 (en) Performing application searches
CN104679743B (en) A kind of method and device of the preference pattern of determining user
CN110532479A (en) A kind of information recommendation method, device and equipment
CN112052387B (en) Content recommendation method, device and computer readable storage medium
CN103678618A (en) Web service recommendation method based on socializing network platform
CA2985257A1 (en) System and method for ranking search results
CN106919575A (en) application program searching method and device
CN111061954B (en) Search result sorting method and device and storage medium
KR20190128246A (en) Searching methods and apparatus and non-transitory computer-readable storage media
CN110275952A (en) News recommended method, device and medium based on user's short-term interest
Zhou et al. Relevance feature mapping for content-based multimedia information retrieval
US20130346385A1 (en) System and method for a purposeful sharing environment
CN106919588A (en) A kind of application program search system and method
CN106919576A (en) Using the method and device of two grades of classes keywords database search for application now
De Boom et al. Semantics-driven event clustering in Twitter feeds
CN111177559A (en) Text travel service recommendation method and device, electronic equipment and storage medium
CN110795613B (en) Commodity searching method, device and system and electronic equipment
CN111488385A (en) Data processing method and device based on artificial intelligence and computer equipment
CN108133058A (en) A kind of video retrieval method
CN108984711A (en) A kind of personalized APP recommended method based on layering insertion
KR100987330B1 (en) A system and method generating multi-concept networks based on user's web usage data

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
CB02 Change of applicant information

Address after: 100085 information Road No. 2, Haidian District, Beijing (Beijing 1-8, D 2-2), two storey 207A room.

Applicant after: QILIN HESHENG NETWORK TECHNOLOGY Inc.

Address before: 100085 information Road No. 2, Haidian District, Beijing (Beijing 1-8, D 2-2), two storey 207A room.

Applicant before: QILIN HESHENG NETWORK TECHNOLOGY Inc.

CB02 Change of applicant information
GR01 Patent grant
GR01 Patent grant