WO2018121700A1 - Method and device for recommending application information based on installed application, terminal device, and storage medium - Google Patents

Method and device for recommending application information based on installed application, terminal device, and storage medium Download PDF

Info

Publication number
WO2018121700A1
WO2018121700A1 PCT/CN2017/119610 CN2017119610W WO2018121700A1 WO 2018121700 A1 WO2018121700 A1 WO 2018121700A1 CN 2017119610 W CN2017119610 W CN 2017119610W WO 2018121700 A1 WO2018121700 A1 WO 2018121700A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
user
application
users
similarity
Prior art date
Application number
PCT/CN2017/119610
Other languages
French (fr)
Chinese (zh)
Inventor
潘岸腾
Original Assignee
广州优视网络科技有限公司
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 广州优视网络科技有限公司 filed Critical 广州优视网络科技有限公司
Publication of WO2018121700A1 publication Critical patent/WO2018121700A1/en

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • G06Q30/0623Item investigation
    • G06Q30/0625Directed, with specific intent or strategy
    • G06Q30/0627Directed, with specific intent or strategy using item specifications
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • G06Q30/0631Item recommendations

Definitions

  • the present application relates to the field of information processing technologies, and in particular, to a method, an apparatus, a terminal device, and a storage medium for recommending application information based on an installed application.
  • the embodiment of the present application provides a method for recommending application information based on an installed application, including:
  • a certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
  • the user is determined according to the number of users who have clicked the information j among the users who have installed the application i and the number of users who have installed the application i.
  • the similarity between the application i and the information j in the stream library has been installed.
  • the similarity is calculated using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • N is an integer greater than zero.
  • the matching degree of the user to the information j is determined according to the number of days the user last installed the application i and maintaining the today and the obtained similarity.
  • the degree of similarity obtained is the similarity between the installed application i and the information j in the information flow library.
  • the matching degree is calculated using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • the user's interest rate i i for the application i is calculated as follows:
  • n the number of applications the user has installed in the N days until today
  • N is an integer greater than zero.
  • the corresponding information is selected from the candidate information according to a predetermined information quality rule and recommended to the user.
  • the comprehensive quality index values calculated by the three parameters are selected according to the order of the comprehensive quality index values from the largest to the smallest.
  • the calculation method of the comprehensive quality index value is:
  • m represents the amount of information in the information flow library
  • c j represents the number of users who clicked on the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • g j represents the number of users who liked the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • d j represents the number of users who generate the application behavior by the information j;
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the stream library.
  • the embodiment of the present application further provides an apparatus for recommending application information based on an installed application, including:
  • a similarity determining unit configured to determine a similarity between the installed information of the application and the different information in the information flow library
  • a matching degree determining unit configured to use the obtained similarity degree to obtain a matching degree of the user to different information
  • Selecting units for selecting a certain amount of information as candidate information in order of matching degree from large to small are selected from large to small;
  • a recommendation unit for recommending candidate information to the user in descending order of matching degree is a recommendation unit for recommending candidate information to the user in descending order of matching degree.
  • the similarity determining unit determines the similarity between the user j installed information and the information j in the information flow library according to the number of users who clicked the information j among the users who have installed the application i and the number of users who have installed the application i. .
  • the similarity determining unit calculates the similarity using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • N is an integer greater than zero.
  • the matching degree determining unit determines the matching degree of the user to the information j according to the number of days the user last installed the application i and maintains today and the obtained similarity, wherein the obtained similarity is the installed application i and the information.
  • the matching degree determining unit calculates the matching degree using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • the user's interest rate i i for the application i is calculated as follows:
  • n the number of applications the user has installed in the N days until today
  • N is an integer greater than zero.
  • the recommendation unit may be further configured to: select, according to a predetermined information quality rule, the corresponding information from the candidate information to recommend to the user.
  • the recommendation unit according to one of a click rate, a click rate and a conversion rate of the candidate information, or a combination of any two or two parameters, or a comprehensive quality index value calculated according to the three parameters, according to the comprehensive quality
  • the index values are selected from the largest to the smallest, and the corresponding information is recommended to the user.
  • the calculation method of the comprehensive quality index value is:
  • m represents the amount of information in the information flow library
  • c j represents the number of users who clicked on the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • g j represents the number of users who liked the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • d j represents the number of users who generate the application behavior by the information j;
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the stream library.
  • the embodiment of the present application further provides a terminal device, including:
  • One or more processors are One or more processors;
  • One or more applications wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to:
  • a certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
  • the user is determined according to the number of users who have clicked the information j among the users who have installed the application i and the number of users who have installed the application i.
  • the similarity between the application i and the information j in the stream library has been installed.
  • the similarity is calculated using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • N is an integer greater than zero.
  • the matching degree of the user to the information j is determined according to the number of days the user last installed the application i and maintaining the today and the obtained similarity.
  • the degree of similarity obtained is the similarity between the installed application i and the information j in the information flow library.
  • the matching degree is calculated using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • the user's interest rate i i for the application i is calculated as follows:
  • n the number of applications the user has installed in the N days until today
  • N is an integer greater than zero.
  • the method further includes: selecting, according to a predetermined information quality rule, the corresponding information from the candidate information to recommend to the user.
  • the comprehensive quality index values calculated by the three parameters are selected according to the order of the comprehensive quality index values from the largest to the smallest.
  • the calculation method of the comprehensive quality index value is:
  • m represents the amount of information in the information flow library
  • c j represents the number of users who clicked on the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • g j represents the number of users who liked the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • d j represents the number of users who generate the application behavior by the information j;
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the stream library.
  • the embodiment of the present application further provides a computer readable storage medium carrying one or more computer instruction programs thereon, when the computer instruction program is executed by one or more processors, the one or more processor implementations are based on
  • the method for installing the application to recommend the application information includes: determining the similarity between the installed information of the user and the different information in the information flow library; using the obtained similarity degree to obtain the matching degree of the user to different information; The order of large to small selects a certain amount of information as candidate information and recommends to the user in order of matching degree from large to small.
  • the method and apparatus for recommending application information based on the installed application according to the present application fully considers the interests and hobbies of the user, and determines the matching of different information by the user by determining the similarity between the installed information and the different information in the information flow library. Degree, thereby selecting a certain number of candidate information, may recommend corresponding candidate information according to the degree of matching from large to small, or may select appropriate information from the candidate information according to the information quality indicator to recommend to the user, which can be based on Different application information of different users' different interests and hobbies, so as to achieve personalized recommendation, which greatly enhances the user experience.
  • FIG. 1A is a screenshot of an example of an existing application store recommendation application
  • FIG. 1B is a screenshot of an example of an existing application store adopting an information flow recommendation application
  • Figure 1C is an example screenshot of a detail page of a message of the information flow
  • FIG. 2 is a flowchart of a method for recommending application information based on an installed application according to the first embodiment of the present application
  • Figure 3 is a screenshot of an example of clicking on a message stream to open a corresponding application
  • FIG. 4 is a schematic block diagram of an apparatus for recommending application information based on an installed application according to a second embodiment of the present application
  • FIG. 5 is a block diagram showing the internal structure of a terminal device according to a third embodiment of the present application.
  • FIG. 2 is a flowchart of a method for recommending application information based on an installed application according to the first embodiment of the present application. Assume that a user A clicks on the information flow section of the application store on the terminal that he uses, and needs to recommend a batch of application information to the user A. The recommended application based on the installed application of the present application as shown in FIG. 2 can be used.
  • a method of information the method comprising the steps of:
  • S1 Determine the similarity between the user's installed application and the different information in the information flow library.
  • the installed application here refers to an application that has been installed on the terminal used by the user when recommending the application information to the user.
  • the first step is to determine the similarity between the user's installed application and the different information in the repository.
  • the specific implementation method is as follows:
  • Each application in the app store or app market will contain 1-3 or more tags, which are used to identify the type to which the app belongs, so that the tag of the app can be used to determine the user's interest, ie by tag identification.
  • the user has installed the category of the app to determine the user's interest.
  • the label that the information is promoted by the application can also be used as the type to which the information belongs, as shown in FIG. 3, for example, the application "Himalaya” which is promoted by opening the information, and the application "Himalaya” can be seen.
  • the details of the bottom position are the two tabs "News Reading” and "Listening" of the app.
  • the application is based on whether the method has the same label, and the application information that is highly correlated is found according to the installed application of the user, and then the similarity between the installed information of the application and the different information in the information flow library is calculated.
  • the user has installed the application here, including the application downloaded and installed by the user through the application store or the application market, and the application that can be found in the application store or the application market.
  • the information flow library is preset when the application development store or application market, and the information in the library can be updated from time to time.
  • Having the same label means that the set of labels included in one application i has an intersection with the set of labels contained in another information j.
  • the similarity between the user j installed information and the information j in the information flow library is determined according to the number of users b i, j who have clicked the information j among the users who have installed the application i and the number of users a i have installed the application i. .
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • N is an integer greater than zero.
  • the meaning of the formula is that when the installed application i and the information j do not have the same label, the similarity is 0; when there is the same label, the similarity is the number of users who clicked the information j among the users who have installed the application i The proportion of the number of users who have installed the application i. The larger the ratio, the higher the similarity.
  • n means that the number of applications that a user keeps installing within N days until today means that, as described above, when a batch of application information is to be recommended to a certain user A, the method is first executed to find recommended information.
  • the number of applications installed on the terminal that the user A is using is counted.
  • the number of installed installed applications is the number of all applications installed on the terminal, preferably refers to the third-party application; of course, N can also be set to a specific limited number of days, for example, 60 days. 90 days and so on.
  • the matching degree of the user A to different information is determined from the multiple applications. In order to find information with high matching.
  • the matching degree of the user to the information j is determined according to the last time the user installs the application i and maintains the number of days t i and the obtained similarity s i,j , wherein the obtained similarity is the installed application i and the information The similarity between the information j in the stream library.
  • the degree of interest o i of the user A to the application i can be obtained according to t i , and then the degree of matching of the user to the information j is determined according to the product of the similarity s i,j and the degree of interest o i .
  • n the number of applications that User A maintains installed within N days until today;
  • m represents the amount of information in the information flow library
  • N is an integer greater than zero.
  • the meaning of the formula is: multiplying the user's interest in the installed application i by the similarity between the application i and the information, and then determining the matching degree of the user to different information by superimposing.
  • n is the same as n in step S1, and no repetitive explanation is made.
  • the user A's interest in the application i is calculated as follows:
  • n the number of applications that User A maintains installed within N days until today;
  • N is an integer greater than 0, which represents the number of days.
  • the meaning of interest o i is that the user's interest will change with time. The closer the user installs the application, the more the user's current interest.
  • n is the same as n in step S1, and no repetitive explanation is made.
  • t i indicates that the user A last installed the application i and maintained the number of days until today means that the user A has been installed on the day when the application i was last installed and has not been uninstalled in the middle but remains in the terminal that the user is using. The number of days until today.
  • S3 Select a certain amount of information as candidate information according to the order of matching degree, and recommend the user to the user in descending order of matching degree.
  • a certain amount of information ranked in front is selected as the candidate information of interest of the user according to the size of the matching degree value. For example, select the corresponding information with the top 50 matching ranks, or the corresponding information ranked in the top 100. The corresponding candidate information can then be recommended to the user in descending order of matching.
  • the method for recommending application information based on the installed application of the present application may further include the following steps:
  • S4 Select corresponding information from the candidate information according to a predetermined information quality rule to recommend to the user.
  • the predetermined information quality rule may be a comprehensive quality indicator value for determining the information, and the comprehensive quality indicator may be one of a click rate, a click rate, and a conversion rate of the candidate information, or may be based on a click rate, a click rate, and a conversion rate.
  • the comprehensive quality index value of the candidate information calculated by any two or two combinations, or the comprehensive quality index value of the candidate information calculated according to the three parameters of click rate, click rate and conversion rate, according to the comprehensive quality index value from large to The small order selects the corresponding information to recommend to the user.
  • the comprehensive quality index value is calculated according to the comprehensive quality index value.
  • the order of the large to small selection of the corresponding information is recommended to the user, wherein the calculation method of the comprehensive quality indicator value is:
  • m represents the amount of information in the information flow library
  • Parameter indicators to calculate the comprehensive quality indicator value are used to calculate the comprehensive quality indicator value;
  • c j represents the number of users who clicked on the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • g j represents the number of users who liked the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • d j represents the number of users who generate the application behavior by the information j;
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the stream library.
  • the above formula considers the click rate, the like rate, and the conversion rate of candidate information as factors for measuring the quality of information.
  • the weights of each factor are adjusted by ⁇ , ⁇ , and ⁇ to consider which one or two factors are considered according to practical needs. As a major measure of quality. The following is a detailed description by way of example.
  • the comprehensive quality index of the candidate information is the click rate of the information.
  • the comprehensive quality index of the candidate information is the click rate of the information.
  • the method for recommending application information based on the installed application according to the present application fully considers the interests and hobbies of the user, and determines the degree of matching of different information by the user by determining the similarity between the installed information and the different information in the information flow library, Therefore, a certain number of candidate information is selected, and the corresponding candidate information may be recommended to the user according to the matching degree from the largest to the smallest, or the appropriate information may be selected from the candidate information according to the information quality indicator, and the user may be recommended according to different users.
  • Different application hobbies and different recommended application information to achieve personalized recommendations, which greatly enhances the user experience.
  • FIG. 4 is a schematic block diagram of an apparatus for recommending application information based on an installed application according to a second embodiment of the present application. As shown in FIG. 4, the apparatus for recommending application information based on an installed application of the present application includes:
  • a similarity determining unit configured to determine a similarity between the installed information of the application and the different information in the information flow library
  • a matching degree determining unit configured to use the obtained similarity degree to obtain a matching degree of the user to different information
  • a recommendation unit for recommending candidate information to the user in descending order of matching degree is a recommendation unit for recommending candidate information to the user in descending order of matching degree.
  • the similarity determination unit determines that the user has installed the application i and the information j in the information flow library according to the number of users who have clicked the information j among the users who have installed the application i and the number of users who have installed the application i. Similarity.
  • the similarity determining unit calculates the similarity using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • N is an integer greater than zero.
  • the matching degree determining unit determines the matching degree of the user to the information j according to the number of days the user last installed the application i and maintains to today and the obtained similarity, wherein the obtained similarity is the installed application i The similarity between the information j and the information j in the information flow library.
  • the matching degree determining unit calculates the matching degree using the following formula:
  • n the number of applications the user has installed in the N days until today
  • m represents the amount of information in the information flow library
  • the degree of user interest for the application of i o i is calculated as follows:
  • n the number of applications the user has installed in the N days until today
  • N is an integer greater than zero.
  • the recommendation unit in the device for recommending application information based on the installed application of the present application may also be used according to
  • the predetermined information quality rule selects the corresponding information from the candidate information and recommends it to the user.
  • the recommendation unit according to one of the click rate, the click rate and the conversion rate of the candidate information, or according to any two or two parameter combinations, or the comprehensive quality index value calculated according to the three parameters,
  • the comprehensive quality index values are selected from the largest to the smallest, and the corresponding information is recommended to the user.
  • the calculation method of the comprehensive quality index value is:
  • m represents the amount of information in the information flow library
  • Parameter indicators to calculate the comprehensive quality indicator value are used to calculate the comprehensive quality indicator value;
  • c j represents the number of users who clicked on the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • g j represents the number of users who liked the information j
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the information flow library
  • d j represents the number of users who generate the application behavior by the information j;
  • f j represents the number of all users who have presented the information j to the user
  • m represents the amount of information in the stream library.
  • the device for recommending application information based on the installed application according to the present application fully considers the interests and hobbies of the user, and determines the degree of matching of the different information by the user by determining the similarity between the installed information and the different information in the information flow library, Therefore, a certain number of candidate information is selected, and the corresponding candidate information may be recommended to the user according to the matching degree from the largest to the smallest, or the appropriate information may be selected from the candidate information according to the information quality indicator, and the user may be recommended according to different users. Different application hobbies and different recommended application information to achieve personalized recommendations, which greatly enhances the user experience.
  • the present application provides a terminal device as a third embodiment, which specifically includes:
  • One or more processors are One or more processors;
  • One or more applications wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to:
  • a certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
  • the terminal device includes a processor 310, a memory 320, an internal memory 330, a network interface 340, and a display screen 350 connected through a system bus.
  • the processor 310 is configured to implement a computing function and a function of controlling the operation of the terminal device, and the processor 310 is configured to perform the method for recommending application information based on the installed application provided by the above embodiment.
  • the processor 310 is configured to determine the similarity between the installed information of the application and the different information in the information flow library; use the obtained similarity to obtain the matching degree of the user to different information; and select a certain order according to the matching degree from large to small The amount of information is used as candidate information and is recommended to the user in descending order of matching.
  • the memory 320 is a non-volatile storage medium storing an operating system 321, a database 322, and a computer program for implementing the method for recommending application information based on the installed application provided by the above embodiment, and a candidate for executing the computer program generation Intermediate data and result data.
  • Network interface 340 is used to communicate with the server, and network interface 340 includes a radio frequency transceiver.
  • the present application also provides a computer readable storage medium carrying one or more computer instruction programs thereon, the one or more processors executing one or more processors executing one
  • the method for recommending application information based on the installed application includes: determining the similarity between the installed information of the user and the different information in the information flow library; using the obtained similarity degree to obtain the matching degree of the user to different information; A certain amount of information is selected as candidate information in descending order and recommended to the user in descending order of matching.
  • the foregoing program may be stored in a computer readable storage medium, and the program is executed when executed.
  • the foregoing storage medium includes: a mobile storage device, a random access memory (RAM), a read-only memory (ROM), a magnetic disk, or an optical disk.
  • RAM random access memory
  • ROM read-only memory
  • magnetic disk or an optical disk.
  • optical disk A medium that can store program code.
  • the above-described integrated unit of the present application may be stored in a computer readable storage medium if it is implemented in the form of a software function module and sold or used as a stand-alone product.
  • the technical solution of the embodiments of the present application may be embodied in the form of a software product in essence or in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for making
  • a computer device which may be a personal computer, server, or network device, etc.
  • the foregoing storage medium includes various media that can store program codes, such as a mobile storage device, a RAM, a ROM, a magnetic disk, or an optical disk.
  • the method and apparatus for recommending application information based on an installed application according to the present application fully considers the interests and hobbies of the user, and determines the similarity of the different information in the installed application and the information flow library, and then determines the user's different information.
  • Matching degree thereby selecting a certain number of candidate information, may recommend corresponding candidate information to the user according to the matching degree from large to small, or may select appropriate information from the candidate information according to the information quality indicator to recommend to the user, which can Different application information recommended according to different users' interests and hobbies, thereby achieving personalized recommendation, improving the viewing rate of application information and the usage rate of the application.

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The present application provides a method and device for recommending application information based on an installed application. The method comprises: determining the similarity between an application installed by a user and different information in an information flow library; obtaining degrees of matching between the user and the different information using the obtained similarity; and selecting a certain amount of information as candidate information according to an order of descending degrees of matching and recommending the candidate information to the user according to the order of descending degrees of matching.

Description

基于已安装应用来推荐应用信息的方法、装置、终端设备及存储介质Method, device, terminal device and storage medium for recommending application information based on installed application
交互参考Cross reference
本申请要求以下优先权:2016年12月29日提出的申请号:201611242501.6,名称:“一种基于已安装应用来推荐应用信息的方法和装置”的中国专利,本申请参考引用了如上所述申请的全部内容。The present application claims the following priority: Application No.: 201611242501.6, filed on Dec. 29, 2016, entitled: "A Method and Apparatus for Recommending Application Information Based on Installed Applications", which is incorporated herein by reference. The entire content of the application.
技术领域Technical field
本申请涉及信息处理技术领域,具体而言涉及一种基于已安装应用来推荐应用信息的方法、装置、终端设备及存储介质。The present application relates to the field of information processing technologies, and in particular, to a method, an apparatus, a terminal device, and a storage medium for recommending application information based on an installed application.
背景技术Background technique
随着互联网技术和智能移动终端技术的快速发展,很多在计算机终端上实现的功能(例如购物、阅读)也都可以在智能移动终端上实现,例如使用智能手机或平板电脑等。另外,这些功能的实现需要在智能移动终端上安装相应的应用程序。例如,网上购物,需要安装例如淘宝客户端,听音乐需要安装音乐播放器客户端等。由此,很多软件公司提供了应用商店或应用市场,例如豌豆荚或者PP助手等。用户可以打开应用商店或者应用市场,从而能够快速搜索和下载所需要的各种应用程序,包括影音播放类、系统工具类、通讯社交类、网上购物类、阅读类等,当然还可以下载游戏等休闲娱乐类应用程序(APP)。With the rapid development of Internet technologies and smart mobile terminal technologies, many functions implemented on computer terminals (such as shopping, reading) can also be implemented on smart mobile terminals, such as using a smart phone or a tablet. In addition, the implementation of these functions requires the installation of the corresponding application on the smart mobile terminal. For example, online shopping requires installing a Taobao client, for example, to listen to music, and to install a music player client. As a result, many software companies offer application stores or application markets, such as pea pods or PP assistants. Users can open the app store or the app market, so they can quickly search and download the various applications they need, including video playback, system tools, communication and social, online shopping, reading, etc. Of course, you can download games, etc. Entertainment app (APP).
为了不断提升用户使用应用商店或者应用市场的良好体验感,目前开发商开发出很多便捷用户使用的功能,其中之一是推荐功能,即向用户推荐一些应用,以帮助用户发现更多感兴趣的应用。对于如何给用户展示应用,传统的做法是把应用直接展示给用户,如图1A所示。这种做法直接了当,但存在一个严重缺陷:缺乏对所推荐应用的介绍。当用户看到一款未 知的应用时,由于没有详细的介绍,绝大多数用户都会因为对它缺乏了解而不产生点击和下载行为。为了解决这个问题,目前出现了一种新的应用发行方式:在应用商店增加信息流,通过有趣的文章、视频对应用进行介绍和推销,如图1B和图1C所示。In order to continuously improve the user experience of using the app store or the application market, developers have developed a number of convenient user-use functions, one of which is the recommendation function, which is to recommend some applications to users to help users find more interesting. application. The traditional approach to how to present an application to a user is to present the application directly to the user, as shown in Figure 1A. This approach is straightforward, but there is a serious flaw: the lack of an introduction to the recommended application. When a user sees an unknown application, because there is no detailed introduction, most users will not click and download because they lack understanding of it. In order to solve this problem, a new application distribution method has emerged: adding information flow in the application store, introducing and promoting the application through interesting articles and videos, as shown in FIG. 1B and FIG. 1C.
但是,对于这种应用商店通过信息流发行应用的方式,在给用户推荐信息过程中,由于不同的用户具有不同的兴趣,所以需要考虑对不同用户推荐不同的应用信息,这就对有针对性的精准推荐提出了很高要求,否则向用户展示的不是他感兴趣的应用,将会大大降低用户的使用体验感。However, for such an application store to distribute applications through information flow, in the process of recommending information to users, since different users have different interests, it is necessary to consider recommending different application information to different users, which is targeted. The precise recommendation puts forward high requirements, otherwise it will not show the user the application that he is interested in, which will greatly reduce the user experience.
申请内容Application content
本申请的目的在于提供一种基于已安装应用来推荐应用信息的方法和装置,以改善上述问题。It is an object of the present application to provide a method and apparatus for recommending application information based on an installed application to improve the above problems.
本申请实施例提供了一种基于已安装应用来推荐应用信息的方法,包括:The embodiment of the present application provides a method for recommending application information based on an installed application, including:
确定用户已安装应用与信息流库里的不同信息的相似度;Determine the similarity between the user's installed application and the different information in the stream library;
利用所获得的相似度来得到用户对不同信息的匹配度;以及Using the obtained similarity to obtain the user's matching degree to different information;
按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。A certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
其中,在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。Wherein, in the step of determining the similarity between the user and the different information in the information flow library, the user is determined according to the number of users who have clicked the information j among the users who have installed the application i and the number of users who have installed the application i. The similarity between the application i and the information j in the stream library has been installed.
其中,在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,使用下列公式计算所述相似度:Wherein, in the step of determining the similarity between the user and the different information in the information flow library, the similarity is calculated using the following formula:
Figure PCTCN2017119610-appb-000001
Figure PCTCN2017119610-appb-000001
其中:among them:
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
N为大于0的整数。N is an integer greater than zero.
其中,在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。Wherein, in the step of using the obtained similarity to obtain the matching degree of the user to different information, the matching degree of the user to the information j is determined according to the number of days the user last installed the application i and maintaining the today and the obtained similarity. The degree of similarity obtained is the similarity between the installed application i and the information j in the information flow library.
其中,在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,使用下列公式计算所述匹配度:Wherein, in the step of using the obtained similarity to obtain the degree of matching of the user to different information, the matching degree is calculated using the following formula:
Figure PCTCN2017119610-appb-000002
Figure PCTCN2017119610-appb-000002
其中:among them:
u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述的用户对应用i的兴趣度o i的计算方法如下: The user's interest rate i i for the application i is calculated as follows:
Figure PCTCN2017119610-appb-000003
Figure PCTCN2017119610-appb-000003
t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
N为大于0的整数。N is an integer greater than zero.
其中,根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。Wherein, the corresponding information is selected from the candidate information according to a predetermined information quality rule and recommended to the user.
其中,在根据预定的信息质量规则从候选信息里选取相应信息向用户推荐的步骤中,根据候选信息的点击率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。Wherein, in the step of recommending the corresponding information from the candidate information to the user according to the predetermined information quality rule, according to one of the click rate, the click rate and the conversion rate of the candidate information, or according to any two or two parameter combinations, or according to the The comprehensive quality index values calculated by the three parameters are selected according to the order of the comprehensive quality index values from the largest to the smallest.
其中,所述综合质量指标值的计算方法为:Wherein, the calculation method of the comprehensive quality index value is:
qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
其中:among them:
qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算所述综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. a parameter indicator to calculate the comprehensive quality indicator value;
所述信息j的点击率
Figure PCTCN2017119610-appb-000004
The click rate of the information j
Figure PCTCN2017119610-appb-000004
c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述信息j的点赞率
Figure PCTCN2017119610-appb-000005
The rate of praise of the information j
Figure PCTCN2017119610-appb-000005
g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述信息j的转化率
Figure PCTCN2017119610-appb-000006
Conversion rate of the information j
Figure PCTCN2017119610-appb-000006
d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量。m represents the amount of information in the stream library.
本申请实施例还提供了一种基于已安装应用来推荐应用信息的装置,包括:The embodiment of the present application further provides an apparatus for recommending application information based on an installed application, including:
相似度确定单元,用于确定用户已安装应用与信息流库里的不同信息的相似度;a similarity determining unit, configured to determine a similarity between the installed information of the application and the different information in the information flow library;
匹配度确定单元,用于利用所获得的相似度来得到用户对不同信息的匹配度;a matching degree determining unit, configured to use the obtained similarity degree to obtain a matching degree of the user to different information;
选取单元,用于按照匹配度从大到小的顺序选取一定数量的信息作为候选信息;以及Selecting units for selecting a certain amount of information as candidate information in order of matching degree from large to small;
推荐单元,用于按匹配度从大到小顺序向用户推荐候选信息。A recommendation unit for recommending candidate information to the user in descending order of matching degree.
其中,所述相似度确定单元根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。The similarity determining unit determines the similarity between the user j installed information and the information j in the information flow library according to the number of users who clicked the information j among the users who have installed the application i and the number of users who have installed the application i. .
其中,所述相似度确定单元使用下列公式计算所述相似度:Wherein the similarity determining unit calculates the similarity using the following formula:
Figure PCTCN2017119610-appb-000007
Figure PCTCN2017119610-appb-000007
其中:among them:
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i= 1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i= 1,2,...,n j=1,2...,m,
a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
N为大于0的整数。N is an integer greater than zero.
其中,所述匹配度确定单元根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。The matching degree determining unit determines the matching degree of the user to the information j according to the number of days the user last installed the application i and maintains today and the obtained similarity, wherein the obtained similarity is the installed application i and the information. The similarity between the information j in the stream library.
其中,所述匹配度确定单元使用下列公式计算所述匹配度:Wherein the matching degree determining unit calculates the matching degree using the following formula:
Figure PCTCN2017119610-appb-000008
Figure PCTCN2017119610-appb-000008
其中:among them:
u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述的用户对应用i的兴趣度o i的计算方法如下: The user's interest rate i i for the application i is calculated as follows:
Figure PCTCN2017119610-appb-000009
Figure PCTCN2017119610-appb-000009
t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
N为大于0的整数。N is an integer greater than zero.
其中,所述推荐单元还可以用于根据预定的信息质量规则从候选信息 里选取相应信息向用户推荐。The recommendation unit may be further configured to: select, according to a predetermined information quality rule, the corresponding information from the candidate information to recommend to the user.
其中,在所述推荐单元中,根据候选信息的点击率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。Wherein, in the recommendation unit, according to one of a click rate, a click rate and a conversion rate of the candidate information, or a combination of any two or two parameters, or a comprehensive quality index value calculated according to the three parameters, according to the comprehensive quality The index values are selected from the largest to the smallest, and the corresponding information is recommended to the user.
其中,所述综合质量指标值的计算方法为:Wherein, the calculation method of the comprehensive quality index value is:
qul i=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul i =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
其中:among them:
qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算所述综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. a parameter indicator to calculate the comprehensive quality indicator value;
所述信息j的点击率
Figure PCTCN2017119610-appb-000010
The click rate of the information j
Figure PCTCN2017119610-appb-000010
c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述信息j的点赞率
Figure PCTCN2017119610-appb-000011
The rate of praise of the information j
Figure PCTCN2017119610-appb-000011
g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述信息j的转化率
Figure PCTCN2017119610-appb-000012
Conversion rate of the information j
Figure PCTCN2017119610-appb-000012
d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量。m represents the amount of information in the stream library.
本申请实施例提供还提供一种终端设备,包括:The embodiment of the present application further provides a terminal device, including:
一个或多个处理器;One or more processors;
存储器;Memory
一个或多个应用程序,其中所述一个或多个应用程序被存储在所述存储器中并被配置为由所述一个或多个处理器执行,所述一个或多个程序配置用于:One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to:
确定用户已安装应用与信息流库里的不同信息的相似度;Determine the similarity between the user's installed application and the different information in the stream library;
利用所获得的相似度来得到用户对不同信息的匹配度;以及Using the obtained similarity to obtain the user's matching degree to different information;
按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。A certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
其中,在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。Wherein, in the step of determining the similarity between the user and the different information in the information flow library, the user is determined according to the number of users who have clicked the information j among the users who have installed the application i and the number of users who have installed the application i. The similarity between the application i and the information j in the stream library has been installed.
其中,在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,使用下列公式计算所述相似度:Wherein, in the step of determining the similarity between the user and the different information in the information flow library, the similarity is calculated using the following formula:
Figure PCTCN2017119610-appb-000013
Figure PCTCN2017119610-appb-000013
其中:among them:
s i,表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i, indicating the similarity between the installed application i and the information j, i=1, 2, ..., n j=1, 2..., m,
b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
N为大于0的整数。N is an integer greater than zero.
其中,在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。Wherein, in the step of using the obtained similarity to obtain the matching degree of the user to different information, the matching degree of the user to the information j is determined according to the number of days the user last installed the application i and maintaining the today and the obtained similarity. The degree of similarity obtained is the similarity between the installed application i and the information j in the information flow library.
其中,在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,使用下列公式计算所述匹配度:Wherein, in the step of using the obtained similarity to obtain the degree of matching of the user to different information, the matching degree is calculated using the following formula:
Figure PCTCN2017119610-appb-000014
Figure PCTCN2017119610-appb-000014
其中:among them:
u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述的用户对应用i的兴趣度o i的计算方法如下: The user's interest rate i i for the application i is calculated as follows:
Figure PCTCN2017119610-appb-000015
Figure PCTCN2017119610-appb-000015
t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
N为大于0的整数。N is an integer greater than zero.
其中,还包括:根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。The method further includes: selecting, according to a predetermined information quality rule, the corresponding information from the candidate information to recommend to the user.
其中,在根据预定的信息质量规则从候选信息里选取相应信息向用户推荐的步骤中,根据候选信息的点击率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。Wherein, in the step of recommending the corresponding information from the candidate information to the user according to the predetermined information quality rule, according to one of the click rate, the click rate and the conversion rate of the candidate information, or according to any two or two parameter combinations, or according to the The comprehensive quality index values calculated by the three parameters are selected according to the order of the comprehensive quality index values from the largest to the smallest.
其中,所述综合质量指标值的计算方法为:Wherein, the calculation method of the comprehensive quality index value is:
qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
其中:among them:
qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算所述综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. a parameter indicator to calculate the comprehensive quality indicator value;
所述信息j的点击率
Figure PCTCN2017119610-appb-000016
The click rate of the information j
Figure PCTCN2017119610-appb-000016
c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述信息j的点赞率
Figure PCTCN2017119610-appb-000017
The rate of praise of the information j
Figure PCTCN2017119610-appb-000017
g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
所述信息j的转化率
Figure PCTCN2017119610-appb-000018
Conversion rate of the information j
Figure PCTCN2017119610-appb-000018
d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量。m represents the amount of information in the stream library.
本申请实施例还提供一种计算机可读存储介质,其上承载一个或多个计算机指令程序,所述计算机指令程序被一个或多个处理器执行时,所述一个或多个处理器实现基于已安装应用来推荐应用信息的方法,包括:确定用户已安装应用与信息流库里的不同信息的相似度;利用所获得的相似度来得到用户对不同信息的匹配度;以及按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。The embodiment of the present application further provides a computer readable storage medium carrying one or more computer instruction programs thereon, when the computer instruction program is executed by one or more processors, the one or more processor implementations are based on The method for installing the application to recommend the application information includes: determining the similarity between the installed information of the user and the different information in the information flow library; using the obtained similarity degree to obtain the matching degree of the user to different information; The order of large to small selects a certain amount of information as candidate information and recommends to the user in order of matching degree from large to small.
根据本申请的基于已安装应用来推荐应用信息的方法和装置充分考虑了用户的兴趣和爱好,通过确定已安装应用与信息流库里的不同信息的相似度,继而确定用户对不同信息的匹配度,从而选取出一定数量的候选信息,可以按匹配度从大到小顺序向用户推荐相应的候选信息,也可以根据信息质量指标从这些候选信息中选取合适的信息向用户推荐,这能够根据不同用户的兴趣爱好不同而推荐的不同应用信息,从而实现个性化推荐,这大大提升了用户的体验感。The method and apparatus for recommending application information based on the installed application according to the present application fully considers the interests and hobbies of the user, and determines the matching of different information by the user by determining the similarity between the installed information and the different information in the information flow library. Degree, thereby selecting a certain number of candidate information, may recommend corresponding candidate information according to the degree of matching from large to small, or may select appropriate information from the candidate information according to the information quality indicator to recommend to the user, which can be based on Different application information of different users' different interests and hobbies, so as to achieve personalized recommendation, which greatly enhances the user experience.
附图说明DRAWINGS
图1A是现有的应用商店推荐应用的一个实例截图;FIG. 1A is a screenshot of an example of an existing application store recommendation application;
图1B是现有的应用商店采用信息流方式推荐应用的一个实例截图;FIG. 1B is a screenshot of an example of an existing application store adopting an information flow recommendation application;
图1C是信息流的一个信息的详情页的一个实例截图;Figure 1C is an example screenshot of a detail page of a message of the information flow;
图2是本申请第一实施例提供的基于已安装应用来推荐应用信息的方法的流程图;2 is a flowchart of a method for recommending application information based on an installed application according to the first embodiment of the present application;
图3是点击信息流的一个信息而打开对应应用的一个实例截图;Figure 3 is a screenshot of an example of clicking on a message stream to open a corresponding application;
图4是本申请第二实施例提供的基于已安装应用来推荐应用信息的装置的示意性框图;以及4 is a schematic block diagram of an apparatus for recommending application information based on an installed application according to a second embodiment of the present application;
图5为本申请第三实施例终端设备的内部结构框图。FIG. 5 is a block diagram showing the internal structure of a terminal device according to a third embodiment of the present application.
具体实施方式detailed description
下面将结合本申请实施例和附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅是本申请一部分实施例,而不是全部的实施例。通常在此处附图中描述和示出的本申请实施例的组件可以以各种不同的配置来布置和设计。因此,以下对在附图中提供的本申请的实施例的详细描述并非旨在限制要求保护的本申请的范围,而是仅仅表示本申请的选定实施例。基于本申请的实施例,本领域技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例,都属于本申请保护的范围。The technical solutions in the embodiments of the present application are clearly and completely described in the following with reference to the embodiments and the accompanying drawings. It is obvious that the described embodiments are only a part of the embodiments of the present application, and not all embodiments. The components of the embodiments of the present application, which are generally described and illustrated in the figures herein, may be arranged and designed in various different configurations. The detailed description of the embodiments of the present application, which is set forth in the claims All other embodiments obtained by a person skilled in the art based on the embodiments of the present application without creative efforts are within the scope of the present application.
第一实施例First embodiment
图2是本申请第一实施例提供的基于已安装应用来推荐应用信息的方法的流程图。假设某个用户A在其使用的终端上点击进入应用商店的信息流版块,此时需要向用户A推荐一批应用信息,可以使用如图2所示的本申请的基于已安装应用来推荐应用信息的方法,该方法包括以下步骤:2 is a flowchart of a method for recommending application information based on an installed application according to the first embodiment of the present application. Assume that a user A clicks on the information flow section of the application store on the terminal that he uses, and needs to recommend a batch of application information to the user A. The recommended application based on the installed application of the present application as shown in FIG. 2 can be used. A method of information, the method comprising the steps of:
S1:确定用户已安装应用与信息流库里的不同信息的相似度。S1: Determine the similarity between the user's installed application and the different information in the information flow library.
这里的已安装应用是指在向用户推荐应用信息时在用户使用的终端上已安装着的应用。The installed application here refers to an application that has been installed on the terminal used by the user when recommending the application information to the user.
通常可以认为,用户使用的例如智能手机或平板电脑或计算机等终端上安装的各种应用,如游戏类、休闲类、办公类等,是该用户感兴趣的应用,而基于用户已安装应用来推荐关联度高的应用信息,继而针对不同用户推荐其感兴趣的应用信息,实现个性化推荐的目的。It is generally considered that various applications installed on a terminal such as a smartphone or a tablet computer or a computer, such as a game class, a casual class, an office class, etc., are applications that the user is interested in, and based on the user installed application. It is recommended to apply information with high relevance, and then recommend application information of interest to different users to achieve personalized recommendation.
那么如何基于用户已安装应用来推荐关联度高的应用信息呢?首先就要确定用户已安装应用与信息流库里的不同信息的相似度。具体实现方法如下:So how do you recommend application information with high relevance based on the user's installed apps? The first step is to determine the similarity between the user's installed application and the different information in the repository. The specific implementation method is as follows:
在应用商店或应用市场中每一个应用都会包含1-3个或更多的标签,这些标签用于标识应用所属的类型,由此可以通过应用具有的标签来判断 用户的兴趣,即通过标签识别用户已安装应用的类别来判断用户的兴趣。对于信息流库中的信息,也可以通过该信息所推销的应用具有的标签作为该信息所属的类型,如图3所示,例如打开信息所推销的应用“喜马拉雅”,可以看到应用“喜马拉雅”的详细信息,底部位置是该应用具有的2个标签“新闻阅读”和“听书”。本申请就是基于是否具有相同的标签的方法,根据用户已安装应用来寻找关联度高的应用信息,然后计算用户已安装应用与信息流库里的不同信息的相似度。Each application in the app store or app market will contain 1-3 or more tags, which are used to identify the type to which the app belongs, so that the tag of the app can be used to determine the user's interest, ie by tag identification. The user has installed the category of the app to determine the user's interest. For the information in the information flow library, the label that the information is promoted by the application can also be used as the type to which the information belongs, as shown in FIG. 3, for example, the application "Himalaya" which is promoted by opening the information, and the application "Himalaya" can be seen. The details of the bottom position are the two tabs "News Reading" and "Listening" of the app. The application is based on whether the method has the same label, and the application information that is highly correlated is found according to the installed application of the user, and then the similarity between the installed information of the application and the different information in the information flow library is calculated.
由上述可知,这里用户已安装应用包括用户通过应用商店或应用市场下载安装的应用和能够在应用商店或应用市场里找得到的应用。信息流库是应用开发商店或应用市场时预置的,库里的信息可以时常更新。As can be seen from the above, the user has installed the application here, including the application downloaded and installed by the user through the application store or the application market, and the application that can be found in the application store or the application market. The information flow library is preset when the application development store or application market, and the information in the library can be updated from time to time.
具有相同标签就是指一个应用i包含的标签的集合与另一个信息j包含的标签的集合具有交集即可。Having the same label means that the set of labels included in one application i has an intersection with the set of labels contained in another information j.
其中,根据在已安装应用i的用户中点击过信息j的用户数量b i,j和已安装应用i的用户数量a i来确定用户已安装应用i与信息流库里的信息j的相似度。 The similarity between the user j installed information and the information j in the information flow library is determined according to the number of users b i, j who have clicked the information j among the users who have installed the application i and the number of users a i have installed the application i. .
其中,可以使用下列公式计算用户已安装应用与信息流库里的不同信息的相似度:Among them, you can use the following formula to calculate the similarity between the user's installed application and the different information in the information flow library:
Figure PCTCN2017119610-appb-000019
Figure PCTCN2017119610-appb-000019
其中:among them:
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
N为大于0的整数。N is an integer greater than zero.
该公式表示的含义是当已安装应用i与信息j没有相同标签时,其相似度为0;当有相同标签时是相似度是在已安装应用i的用户中点击过信息j的用户数量与已安装应用i的用户数量的比例,该比例越大,相似度越高。The meaning of the formula is that when the installed application i and the information j do not have the same label, the similarity is 0; when there is the same label, the similarity is the number of users who clicked the information j among the users who have installed the application i The proportion of the number of users who have installed the application i. The larger the ratio, the higher the similarity.
n表示某个用户在到今天为止的N天内保持安装的应用的数量的意思是指:正如上述那样,当准备向某个用户A推荐一批应用信息时,先执行本方法寻找推荐信息,此时统计该用户A正使用的终端上安装着的应用的数量。当表示天数的N足够大时,统计的已安装应用的数量就是终端上安装的全部应用的数量,优选是指第三方应用;当然也可以将N设定为具体的有限天数,例如60天、90天等。n means that the number of applications that a user keeps installing within N days until today means that, as described above, when a batch of application information is to be recommended to a certain user A, the method is first executed to find recommended information. The number of applications installed on the terminal that the user A is using is counted. When the number of days indicating N is sufficiently large, the number of installed installed applications is the number of all applications installed on the terminal, preferably refers to the third-party application; of course, N can also be set to a specific limited number of days, for example, 60 days. 90 days and so on.
S2:利用所获得的相似度来得到用户对不同信息的匹配度。S2: Using the obtained similarity to obtain the matching degree of the user to different information.
在得到用户已安装应用与信息流库里的不同信息的相似度之后,由于用户使用的终端上常常会安装多个应用,就要从该多个应用中确定该用户A对不同信息的匹配度,以便寻找匹配度高的信息。After obtaining the similarity of the different information in the installed application and the information flow library, since the user often installs multiple applications on the terminal, the matching degree of the user A to different information is determined from the multiple applications. In order to find information with high matching.
其中,根据用户最近一次安装应用i并保持到今天的天数t i和所获得的相似度s i,j来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。例如,可以根据t i得到用户A对应用i的兴趣度o i,然后根据相似度s i,j和兴趣度o i的乘积确定用户对信息j的匹配度。 Wherein, the matching degree of the user to the information j is determined according to the last time the user installs the application i and maintains the number of days t i and the obtained similarity s i,j , wherein the obtained similarity is the installed application i and the information The similarity between the information j in the stream library. For example, the degree of interest o i of the user A to the application i can be obtained according to t i , and then the degree of matching of the user to the information j is determined according to the product of the similarity s i,j and the degree of interest o i .
使用下列公式计算匹配度:Use the following formula to calculate the match:
Figure PCTCN2017119610-appb-000020
Figure PCTCN2017119610-appb-000020
其中:among them:
u j表示用户A对信息j的匹配度,j=1,2…,m, u j represents the matching degree of user A to information j, j=1, 2..., m,
o i表示用户A对应用i的兴趣度,i=1,2,…,n, o i represents user A's interest in application i, i=1, 2,...,n,
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
n表示用户A在到今天为止的N天内保持安装的应用的数量;n represents the number of applications that User A maintains installed within N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
N为大于0的整数。N is an integer greater than zero.
该公式表示的含义是:将用户对已安装应用i的兴趣度乘以该应用i与信息的相似度,然后通过叠加的方式来确定该用户对不同信息的匹配度。The meaning of the formula is: multiplying the user's interest in the installed application i by the similarity between the application i and the information, and then determining the matching degree of the user to different information by superimposing.
这里的n与步骤S1中的n相同,不再做重复性解释。Here, n is the same as n in step S1, and no repetitive explanation is made.
而的用户A对应用i的兴趣度o i的计算方法如下: The user A's interest in the application i is calculated as follows:
Figure PCTCN2017119610-appb-000021
Figure PCTCN2017119610-appb-000021
t i表示用户A最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days that user A installed application i last time and held to today, i=1, 2,...,n;
n表示用户A在到今天为止的N天内保持安装的应用的数量;n represents the number of applications that User A maintains installed within N days until today;
N为大于0的整数,其表示天数。N is an integer greater than 0, which represents the number of days.
兴趣度o i的含义是考虑用户的兴趣会随时间变化,用户安装应用的时间距离现在越近,越能代表用户现在的兴趣。 The meaning of interest o i is that the user's interest will change with time. The closer the user installs the application, the more the user's current interest.
这里的n与步骤S1中的n相同,不再做重复性解释。Here, n is the same as n in step S1, and no repetitive explanation is made.
t i表示用户A最近一次安装应用i并保持到今天的天数的意思是指:从用户A最近一次安装了应用i的那天开始并且中途没有卸载过而是一直保留在用户正使用的终端里,直到今天的天数。 t i indicates that the user A last installed the application i and maintained the number of days until today means that the user A has been installed on the day when the application i was last installed and has not been uninstalled in the middle but remains in the terminal that the user is using. The number of days until today.
S3:按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并 且按匹配度从大到小顺序向用户推荐。S3: Select a certain amount of information as candidate information according to the order of matching degree, and recommend the user to the user in descending order of matching degree.
在得到了用户A对不同信息的匹配度后,根据匹配度值的大小选取排名在前面的一定数量的信息作为用户的感兴趣的候选信息。例如,选取匹配度排名在前50个的相应信息,或者排名在前100个的相应信息等。然后可以按匹配度从大到小顺序向用户推荐相应的候选信息。After the matching degree of the user A to different information is obtained, a certain amount of information ranked in front is selected as the candidate information of interest of the user according to the size of the matching degree value. For example, select the corresponding information with the top 50 matching ranks, or the corresponding information ranked in the top 100. The corresponding candidate information can then be recommended to the user in descending order of matching.
在一个优选实施例中,为了更进一步向用户推荐其感兴趣的应用信息,实现更精确的个性化推荐,本申请的基于已安装应用来推荐应用信息的方法还可以包括步骤:In a preferred embodiment, in order to further recommend the application information of interest to the user to implement more accurate personalized recommendation, the method for recommending application information based on the installed application of the present application may further include the following steps:
S4:根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。S4: Select corresponding information from the candidate information according to a predetermined information quality rule to recommend to the user.
预定的信息质量规则可以是确定信息的综合质量指标值,而综合质量指标可以是候选信息的点击率、点赞率和转化率之一,或者是根据点击率、点赞率和转化率之间的任意两两组合计算出的候选信息的综合质量指标值、或者根据点击率、点赞率和转化率这三个参数计算出的候选信息的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。The predetermined information quality rule may be a comprehensive quality indicator value for determining the information, and the comprehensive quality indicator may be one of a click rate, a click rate, and a conversion rate of the candidate information, or may be based on a click rate, a click rate, and a conversion rate. The comprehensive quality index value of the candidate information calculated by any two or two combinations, or the comprehensive quality index value of the candidate information calculated according to the three parameters of click rate, click rate and conversion rate, according to the comprehensive quality index value from large to The small order selects the corresponding information to recommend to the user.
具体而言,根据候选信息的点击率、点赞率和转化率之一、或者根据它们之间的任意两两组合、或者根据该三个参数计算出综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐,其中综合质量指标值的计算方法为:Specifically, according to one of the click rate, the click rate, and the conversion rate of the candidate information, or according to any two or two combinations between them, or based on the three parameters, the comprehensive quality index value is calculated according to the comprehensive quality index value. The order of the large to small selection of the corresponding information is recommended to the user, wherein the calculation method of the comprehensive quality indicator value is:
qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
其中:among them:
qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. Parameter indicators to calculate the comprehensive quality indicator value;
信息j的点击率
Figure PCTCN2017119610-appb-000022
Information j click rate
Figure PCTCN2017119610-appb-000022
c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
信息j的点赞率
Figure PCTCN2017119610-appb-000023
Information j rating
Figure PCTCN2017119610-appb-000023
g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
信息j的转化率
Figure PCTCN2017119610-appb-000024
Conversion rate of information j
Figure PCTCN2017119610-appb-000024
d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量。m represents the amount of information in the stream library.
上述公式考虑了候选信息的点击率、点赞率和转化率作为衡量信息质量的因素,通过θ、γ和β来调节每个因素的权重,以根据实践需要侧重考虑哪1个或2个因素作为质量的主要衡量因素。下面通过举例来详细说明。The above formula considers the click rate, the like rate, and the conversion rate of candidate information as factors for measuring the quality of information. The weights of each factor are adjusted by θ, γ, and β to consider which one or two factors are considered according to practical needs. As a major measure of quality. The following is a detailed description by way of example.
β+γ+θ=1,且β、γ和θ∈[0,1],当θ=1、γ=0和β=0时,则候选信息的综合质量指标是该信息的点击率,当θ=0、γ=1和β=0时,则候选信息的综合质量指标是该信息的点赞率,当θ=0、γ=0和β=1时,则候选信息的综合质量指标是该信息的转化率;当θ=0.5、γ=0.5和β=0(θ和γ也可以取非0的其它值,且满足θ+γ=1即可)时,候选信息的综合质量指标是该信息的点击率和点赞率的组合考量,也可以让θ=0、γ=0.5和β=0.5(γ和β也可以取非0的 其它值,且满足γ+β=1即可,或者θ=0.5、γ=0和β=0.5(θ和β也可以取非0的其它值,且满足θ+β=1即可),以此类推,还可以选取其它两两组合,这里不再举例;如果将这3种因素同时考虑,则β+γ+θ=1,且β、γ和θ∈(0,1),即β、γ和θ均不为0和1,当认为某个因素作为质量的主要衡量因素时,可以使该因素的权重值更大,例如θ=0.5、γ=0.3和β=0.2,此时认为点击率是作为候选信息的质量指标的主要衡量因素,点赞率其次,转化率排在最后;以此类推,也可以选取不同的β、γ和θ值来列举点击率、点赞率和转化率作为候选信息的质量指标的主要衡量因素的前后顺序,当然也可以使得点击率、点赞率和转化率作为衡量因素同等重要,这里不再举例描述。β+γ+θ=1, and β, γ, and θ∈[0,1], when θ=1, γ=0, and β=0, the comprehensive quality index of the candidate information is the click rate of the information. When θ=0, γ=1 and β=0, the comprehensive quality index of the candidate information is the click rate of the information. When θ=0, γ=0 and β=1, the comprehensive quality index of the candidate information is The conversion rate of the information; when θ = 0.5, γ = 0.5, and β = 0 (θ and γ can also take other values other than 0, and θ + γ = 1 is satisfied), the comprehensive quality index of the candidate information is The combination of the click rate and the like rate of the information may also allow θ=0, γ=0.5, and β=0.5 (γ and β may also take other values other than 0, and satisfy γ+β=1, Or θ=0.5, γ=0, and β=0.5 (θ and β can also take other values other than 0, and satisfy θ+β=1), and so on, and other two-two combinations can be selected. For another example; if these three factors are considered at the same time, β + γ + θ = 1, and β, γ, and θ ∈ (0, 1), that is, β, γ, and θ are not 0 and 1, when When a factor is used as the main measure of quality, the weight value of the factor can be made Larger, such as θ = 0.5, γ = 0.3, and β = 0.2, at this time, the click rate is considered to be the main measure of the quality indicator of the candidate information, followed by the rate of praise, and the conversion rate is ranked last; and so on. Select different β, γ and θ values to enumerate the click-through rate, click rate and conversion rate as the main measure of the quality indicators of the candidate information. Of course, the click rate, the like rate and the conversion rate can be used as the measurement factors. Equally important, no longer described here.
根据本申请的基于已安装应用来推荐应用信息的方法充分考虑了用户的兴趣和爱好,通过确定已安装应用与信息流库里的不同信息的相似度,继而确定用户对不同信息的匹配度,从而选取出一定数量的候选信息,可以按匹配度从大到小顺序向用户推荐相应的候选信息,也可以根据信息质量指标从这些候选信息中选取合适的信息向用户推荐,这能够根据不同用户的兴趣爱好不同而推荐的不同应用信息,从而实现个性化推荐,这大大提升了用户的体验感。The method for recommending application information based on the installed application according to the present application fully considers the interests and hobbies of the user, and determines the degree of matching of different information by the user by determining the similarity between the installed information and the different information in the information flow library, Therefore, a certain number of candidate information is selected, and the corresponding candidate information may be recommended to the user according to the matching degree from the largest to the smallest, or the appropriate information may be selected from the candidate information according to the information quality indicator, and the user may be recommended according to different users. Different application hobbies and different recommended application information to achieve personalized recommendations, which greatly enhances the user experience.
第二实施例Second embodiment
图4是本申请第二实施例提供的基于已安装应用来推荐应用信息的装置的示意性框图。如图4所示,本申请的基于已安装应用来推荐应用信息的装置包括:FIG. 4 is a schematic block diagram of an apparatus for recommending application information based on an installed application according to a second embodiment of the present application. As shown in FIG. 4, the apparatus for recommending application information based on an installed application of the present application includes:
相似度确定单元,用于确定用户已安装应用与信息流库里的不同信息的相似度;a similarity determining unit, configured to determine a similarity between the installed information of the application and the different information in the information flow library;
匹配度确定单元,用于利用所获得的相似度来得到用户对不同信息的匹配度;a matching degree determining unit, configured to use the obtained similarity degree to obtain a matching degree of the user to different information;
选取单元,用于按照匹配度从大到小的顺序选取一定数量的信息作为候选信息;Selecting units for selecting a certain amount of information as candidate information according to the order of matching degree from large to small;
推荐单元,用于按匹配度从大到小顺序向用户推荐候选信息。A recommendation unit for recommending candidate information to the user in descending order of matching degree.
在一些实施例中,相似度确定单元根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。In some embodiments, the similarity determination unit determines that the user has installed the application i and the information j in the information flow library according to the number of users who have clicked the information j among the users who have installed the application i and the number of users who have installed the application i. Similarity.
在一个优选方案中,相似度确定单元使用下列公式计算相似度:In a preferred embodiment, the similarity determining unit calculates the similarity using the following formula:
Figure PCTCN2017119610-appb-000025
Figure PCTCN2017119610-appb-000025
其中:among them:
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
N为大于0的整数。N is an integer greater than zero.
在一些实施例中,匹配度确定单元根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。In some embodiments, the matching degree determining unit determines the matching degree of the user to the information j according to the number of days the user last installed the application i and maintains to today and the obtained similarity, wherein the obtained similarity is the installed application i The similarity between the information j and the information j in the information flow library.
在一个优选方案中,匹配度确定单元使用下列公式计算匹配度:In a preferred embodiment, the matching degree determining unit calculates the matching degree using the following formula:
Figure PCTCN2017119610-appb-000026
Figure PCTCN2017119610-appb-000026
其中:among them:
u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
的用户对应用i的兴趣度o i的计算方法如下: The degree of user interest for the application of i o i is calculated as follows:
Figure PCTCN2017119610-appb-000027
Figure PCTCN2017119610-appb-000027
t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
N为大于0的整数。N is an integer greater than zero.
在一个优选实施例中,为了更进一步向用户推荐其感兴趣的应用信息,实现更精确的个性化推荐,本申请的基于已安装应用来推荐应用信息的装置中的推荐单元还可以用于根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。In a preferred embodiment, in order to further recommend the application information of interest to the user to achieve more accurate personalized recommendation, the recommendation unit in the device for recommending application information based on the installed application of the present application may also be used according to The predetermined information quality rule selects the corresponding information from the candidate information and recommends it to the user.
在一个优选方案中,在推荐单元中,根据候选信息的点击率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐,其中综合质量指标值的计算方法为:In a preferred solution, in the recommendation unit, according to one of the click rate, the click rate and the conversion rate of the candidate information, or according to any two or two parameter combinations, or the comprehensive quality index value calculated according to the three parameters, The comprehensive quality index values are selected from the largest to the smallest, and the corresponding information is recommended to the user. The calculation method of the comprehensive quality index value is:
qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
其中:among them:
qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1], 通过取β、γ和θ不同值来确定通过那些参数指标来计算综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. Parameter indicators to calculate the comprehensive quality indicator value;
信息j的点击率
Figure PCTCN2017119610-appb-000028
Information j click rate
Figure PCTCN2017119610-appb-000028
c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
信息j的点赞率
Figure PCTCN2017119610-appb-000029
Information j rating
Figure PCTCN2017119610-appb-000029
g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
信息j的转化率
Figure PCTCN2017119610-appb-000030
Conversion rate of information j
Figure PCTCN2017119610-appb-000030
d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
m表示信息流库里的信息数量。m represents the amount of information in the stream library.
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的装置的具体工作过程,可以参考前述方法实施例中的对应过程,前述方法实施例中列举的例子和相关描述,同样适用于解释装置的工作过程,在此不再重复描述。A person skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the device described above can refer to the corresponding process in the foregoing method embodiments, the examples and related descriptions in the foregoing method embodiments, The same applies to the working process of the interpretation device, and the description will not be repeated here.
根据本申请的基于已安装应用来推荐应用信息的装置充分考虑了用户的兴趣和爱好,通过确定已安装应用与信息流库里的不同信息的相似度,继而确定用户对不同信息的匹配度,从而选取出一定数量的候选信息,可以按匹配度从大到小顺序向用户推荐相应的候选信息,也可以根据信息质量指标从这些候选信息中选取合适的信息向用户推荐,这能够根据不同用户的兴趣爱好不同而推荐的不同应用信息,从而实现个性化推荐,这大大 提升了用户的体验感。The device for recommending application information based on the installed application according to the present application fully considers the interests and hobbies of the user, and determines the degree of matching of the different information by the user by determining the similarity between the installed information and the different information in the information flow library, Therefore, a certain number of candidate information is selected, and the corresponding candidate information may be recommended to the user according to the matching degree from the largest to the smallest, or the appropriate information may be selected from the candidate information according to the information quality indicator, and the user may be recommended according to different users. Different application hobbies and different recommended application information to achieve personalized recommendations, which greatly enhances the user experience.
第三实施例Third embodiment
为进一步说明第一实施例基于已安装应用来推荐应用信息的方法,本申请提供一终端设备作为第三实施例,具体包括:To further illustrate the method for recommending application information based on the installed application in the first embodiment, the present application provides a terminal device as a third embodiment, which specifically includes:
一个或多个处理器;One or more processors;
存储器;Memory
一个或多个应用程序,其中所述一个或多个应用程序被存储在所述存储器中并被配置为由所述一个或多个处理器执行,所述一个或多个程序配置用于:One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to:
确定用户已安装应用与信息流库里的不同信息的相似度;Determine the similarity between the user's installed application and the different information in the stream library;
利用所获得的相似度来得到用户对不同信息的匹配度;以及Using the obtained similarity to obtain the user's matching degree to different information;
按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。A certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
如图5所示,终端设备包括通过系统总线连接的处理器310、存储器320、内存储器330、网络接口340和显示屏350。处理器310用于实现计算功能和控制终端装置工作的功能,处理器310被配置为执行上述实施例提供的基于已安装应用来推荐应用信息的方法。处理器310用于确定用户已安装应用与信息流库里的不同信息的相似度;利用所获得的相似度来得到用户对不同信息的匹配度;以及按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。存储器320是一种非易失性存储介质,存储有操作系统321、数据库322和用于实现上述实施例提供的基于已安装应用来推荐应用信息的方法的计算机程序,以及执行计算机程序产生的候选中间数据以及结果数据。网络接口340用于与服务器通信,网络接口340包括射频收发器。As shown in FIG. 5, the terminal device includes a processor 310, a memory 320, an internal memory 330, a network interface 340, and a display screen 350 connected through a system bus. The processor 310 is configured to implement a computing function and a function of controlling the operation of the terminal device, and the processor 310 is configured to perform the method for recommending application information based on the installed application provided by the above embodiment. The processor 310 is configured to determine the similarity between the installed information of the application and the different information in the information flow library; use the obtained similarity to obtain the matching degree of the user to different information; and select a certain order according to the matching degree from large to small The amount of information is used as candidate information and is recommended to the user in descending order of matching. The memory 320 is a non-volatile storage medium storing an operating system 321, a database 322, and a computer program for implementing the method for recommending application information based on the installed application provided by the above embodiment, and a candidate for executing the computer program generation Intermediate data and result data. Network interface 340 is used to communicate with the server, and network interface 340 includes a radio frequency transceiver.
第四实施例:Fourth embodiment:
本申请还提供一种计算机可读存储介质,其上承载一个或多个计算机指令程序,所述计算机指令程序被一个或多个处理器执行时,所述一个或多个处理器执行实现一种基于已安装应用来推荐应用信息的方法,包括:确定用户已安装应用与信息流库里的不同信息的相似度;利用所获得的相似度来得到用户对不同信息的匹配度;以及按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。The present application also provides a computer readable storage medium carrying one or more computer instruction programs thereon, the one or more processors executing one or more processors executing one The method for recommending application information based on the installed application includes: determining the similarity between the installed information of the user and the different information in the information flow library; using the obtained similarity degree to obtain the matching degree of the user to different information; A certain amount of information is selected as candidate information in descending order and recommended to the user in descending order of matching.
本领域普通技术人员可以理解:实现上述方法实施例的全部或部分步骤可以通过程序指令相关的硬件来完成,前述的程序可以存储于一计算机可读取存储介质中,该程序在执行时,执行包括上述任意方法实施例的步骤;而前述的存储介质包括:移动存储设备、随机存取存储器(RAM,Random Access Memory)、只读存储器(ROM,Read-Only Memory)、磁碟或者光盘等各种可以存储程序代码的介质。A person skilled in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by using hardware related to the program instructions. The foregoing program may be stored in a computer readable storage medium, and the program is executed when executed. The foregoing storage medium includes: a mobile storage device, a random access memory (RAM), a read-only memory (ROM), a magnetic disk, or an optical disk. A medium that can store program code.
或者,本申请上述集成的单元如果以软件功能模块的形式实现并作为独立的产品销售或使用时,也可以存储在一个计算机可读取存储介质中。基于这样的理解,本申请实施例的技术方案本质上或者说对相关技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机、服务器、或者网络设备等)执行申请各个实施例所述方法的全部或部分。而前述的存储介质包括:移动存储设备、RAM、ROM、磁碟或者光盘等各种可以存储程序代码的介质。Alternatively, the above-described integrated unit of the present application may be stored in a computer readable storage medium if it is implemented in the form of a software function module and sold or used as a stand-alone product. Based on such understanding, the technical solution of the embodiments of the present application may be embodied in the form of a software product in essence or in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for making A computer device (which may be a personal computer, server, or network device, etc.) performs all or part of the methods described in the various embodiments. The foregoing storage medium includes various media that can store program codes, such as a mobile storage device, a RAM, a ROM, a magnetic disk, or an optical disk.
以上所述,仅为本申请的具体实施方式,但本申请的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本申请揭露的技术范围内,可轻易想到变化或替换,都应涵盖在本申请的保护范围之内。因此,本申请的保护范围应以所述权利要求的保护范围为准。The foregoing is only a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application. It should be covered by the scope of protection of this application. Therefore, the scope of protection of the present application should be determined by the scope of the claims.
工业实用性Industrial applicability
根据本申请的基于已安装应用来推荐应用信息的方法和装置,充分考虑了用户的兴趣和爱好,通过确定已安装应用与信息流库里的不同信息的相似度,继而确定用户对不同信息的匹配度,从而选取出一定数量的候选信息,可以按匹配度从大到小顺序向用户推荐相应的候选信息,也可以根据信息质量指标从这些候选信息中选取合适的信息向用户推荐,这能够根据不同用户的兴趣爱好不同而推荐的不同应用信息,从而实现个性化推荐,提高了应用信息的查看率以及应用的使用率。The method and apparatus for recommending application information based on an installed application according to the present application fully considers the interests and hobbies of the user, and determines the similarity of the different information in the installed application and the information flow library, and then determines the user's different information. Matching degree, thereby selecting a certain number of candidate information, may recommend corresponding candidate information to the user according to the matching degree from large to small, or may select appropriate information from the candidate information according to the information quality indicator to recommend to the user, which can Different application information recommended according to different users' interests and hobbies, thereby achieving personalized recommendation, improving the viewing rate of application information and the usage rate of the application.

Claims (25)

  1. 一种基于已安装应用来推荐应用信息的方法,包括:A method for recommending application information based on an installed application, including:
    确定用户已安装应用与信息流库里的不同信息的相似度;Determine the similarity between the user's installed application and the different information in the stream library;
    利用所获得的相似度来得到用户对不同信息的匹配度;以及Using the obtained similarity to obtain the user's matching degree to different information;
    按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。A certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
  2. 根据权利要求1所述的方法,其中在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。The method according to claim 1, wherein in the step of determining that the user has installed the similarity of the application and the different information in the information stream library, the number of users who have clicked the information j among the users who have installed the application i and installed The number of users of i is used to determine the similarity between the user i installed the application i and the information j in the stream library.
  3. 根据权利要求2所述的方法,其中在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,使用下列公式计算所述相似度:The method according to claim 2, wherein in the step of determining that the user has installed the similarity of the application and the different information in the information stream library, the similarity is calculated using the following formula:
    Figure PCTCN2017119610-appb-100001
    Figure PCTCN2017119610-appb-100001
    其中:among them:
    s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
    b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
    a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
    K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
    L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    N为大于0的整数。N is an integer greater than zero.
  4. 根据权利要求1所述的方法,其中在利用所获得的相似度来得到用 户对不同信息的匹配度的步骤中,根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。The method according to claim 1, wherein in the step of obtaining the degree of matching of the user with different information by using the obtained similarity, the number of days and the obtained similarity are maintained according to the last time the user installed the application i and maintained to today. The degree of matching of the user to the information j is determined, wherein the similarity obtained is the similarity between the installed application i and the information j in the information stream library.
  5. 根据权利要求4所述的方法,其中在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,使用下列公式计算所述匹配度:The method according to claim 4, wherein in the step of obtaining the degree of matching of the user with different information by using the obtained similarity, the matching degree is calculated using the following formula:
    Figure PCTCN2017119610-appb-100002
    Figure PCTCN2017119610-appb-100002
    其中:among them:
    u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
    o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
    s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述的用户对应用i的兴趣度o i的计算方法如下: The user's interest rate i i for the application i is calculated as follows:
    Figure PCTCN2017119610-appb-100003
    Figure PCTCN2017119610-appb-100003
    t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    N为大于0的整数。N is an integer greater than zero.
  6. 根据权利要求1所述的方法,其中还包括:根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。The method of claim 1, further comprising: selecting the corresponding information from the candidate information to recommend to the user according to a predetermined information quality rule.
  7. 根据权利要求6所述的方法,其中在根据预定的信息质量规则从候选信息里选取相应信息向用户推荐的步骤中,根据候选信息的点击率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。The method according to claim 6, wherein in the step of recommending the corresponding information from the candidate information to the user according to a predetermined information quality rule, according to one of a click rate, a click rate, and a conversion rate of the candidate information, or according to an arbitrary The combination of the two parameters, or the comprehensive quality index value calculated according to the three parameters, is selected according to the comprehensive quality index value from the largest to the smallest, and recommended to the user.
  8. 根据权利要求7所述的方法,其中所述综合质量指标值的计算方法为:The method of claim 7 wherein said integrated quality indicator value is calculated as:
    qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
    其中:among them:
    qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
    ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
    gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
    dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算所述综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. a parameter indicator to calculate the comprehensive quality indicator value;
    所述信息j的点击率
    Figure PCTCN2017119610-appb-100004
    The click rate of the information j
    Figure PCTCN2017119610-appb-100004
    c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述信息j的点赞率
    Figure PCTCN2017119610-appb-100005
    The rate of praise of the information j
    Figure PCTCN2017119610-appb-100005
    g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述信息j的转化率
    Figure PCTCN2017119610-appb-100006
    Conversion rate of the information j
    Figure PCTCN2017119610-appb-100006
    d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量。m represents the amount of information in the stream library.
  9. 一种基于已安装应用来推荐应用信息的装置,包括:An apparatus for recommending application information based on an installed application, including:
    相似度确定单元,用于确定用户已安装应用与信息流库里的不同信息的相似度;a similarity determining unit, configured to determine a similarity between the installed information of the application and the different information in the information flow library;
    匹配度确定单元,用于利用所获得的相似度来得到用户对不同信息的匹配度;a matching degree determining unit, configured to use the obtained similarity degree to obtain a matching degree of the user to different information;
    选取单元,用于按照匹配度从大到小的顺序选取一定数量的信息作为候选信息;以及Selecting units for selecting a certain amount of information as candidate information in order of matching degree from large to small;
    推荐单元,用于按匹配度从大到小顺序向用户推荐候选信息。A recommendation unit for recommending candidate information to the user in descending order of matching degree.
  10. 根据权利要求9所述的装置,其中所述相似度确定单元根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。The apparatus according to claim 9, wherein said similarity determining unit determines that the user has installed the application i and the information stream based on the number of users who have clicked on the information j among the users who have installed the application i and the number of users who have installed the application i The similarity of the information j in the library.
  11. 根据权利要求10所述的装置,其中所述相似度确定单元使用下列公式计算所述相似度:The apparatus according to claim 10, wherein said similarity determining unit calculates said similarity using the following formula:
    Figure PCTCN2017119610-appb-100007
    Figure PCTCN2017119610-appb-100007
    其中:among them:
    s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
    b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
    a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
    K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
    L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    N为大于0的整数。N is an integer greater than zero.
  12. 根据权利要求9所述的装置,其中所述匹配度确定单元根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。The apparatus according to claim 9, wherein said matching degree determining unit determines the degree of matching of the user to the information j based on the number of days the user last installed the application i and maintained to today and the obtained similarity, wherein the obtained similarity Degree is the similarity between the installed application i and the information j in the information flow library.
  13. 根据权利要求12所述的装置,其中所述匹配度确定单元使用下列公式计算所述匹配度:The apparatus according to claim 12, wherein said matching degree determining unit calculates said matching degree using the following formula:
    Figure PCTCN2017119610-appb-100008
    Figure PCTCN2017119610-appb-100008
    其中:among them:
    u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
    o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
    s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述的用户对应用i的兴趣度o i的计算方法如下: The user's interest rate i i for the application i is calculated as follows:
    Figure PCTCN2017119610-appb-100009
    Figure PCTCN2017119610-appb-100009
    t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    N为大于0的整数。N is an integer greater than zero.
  14. 根据权利要求9所述的装置,其中所述推荐单元还用于根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。The apparatus according to claim 9, wherein said recommending unit is further configured to select corresponding information from the candidate information to recommend the user according to a predetermined information quality rule.
  15. 根据权利要求14所述的装置,其中在所述推荐单元中,根据候选信息的点击率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。The apparatus according to claim 14, wherein in said recommending unit, based on one of a click rate, a click rate, and a conversion rate of the candidate information, or based on any two or two parameter combinations, or calculated based on the three parameters The comprehensive quality index value is selected according to the order of the comprehensive quality index values from the largest to the smallest.
  16. 根据权利要求15所述的装置,其中所述综合质量指标值的计算方法为:The apparatus of claim 15 wherein said integrated quality indicator value is calculated as:
    qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
    其中:among them:
    qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
    ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
    gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
    dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算所述综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. a parameter indicator to calculate the comprehensive quality indicator value;
    所述信息j的点击率
    Figure PCTCN2017119610-appb-100010
    The click rate of the information j
    Figure PCTCN2017119610-appb-100010
    c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述信息j的点赞率
    Figure PCTCN2017119610-appb-100011
    The rate of praise of the information j
    Figure PCTCN2017119610-appb-100011
    g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述信息j的转化率
    Figure PCTCN2017119610-appb-100012
    Conversion rate of the information j
    Figure PCTCN2017119610-appb-100012
    d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量。m represents the amount of information in the stream library.
  17. 一种终端设备,包括:A terminal device comprising:
    一个或多个处理器;One or more processors;
    存储器;Memory
    一个或多个应用程序,其中所述一个或多个应用程序被存储在所述存储器中并被配置为由所述一个或多个处理器执行,所述一个或多个程序配置用于:One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to:
    确定用户已安装应用与信息流库里的不同信息的相似度;Determine the similarity between the user's installed application and the different information in the stream library;
    利用所获得的相似度来得到用户对不同信息的匹配度;以及Using the obtained similarity to obtain the user's matching degree to different information;
    按照匹配度从大到小的顺序选取一定数量的信息作为候选信息并且按匹配度从大到小顺序向用户推荐。A certain amount of information is selected as candidate information in descending order of matching degree and recommended to the user in order of matching degree from large to small.
  18. 根据权利要求17所述的终端设备,其中在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,根据在已安装应用i的用户中点击过信息j的用户数量和已安装应用i的用户数量来确定用户已安装应用i与信息流库里的信息j的相似度。The terminal device according to claim 17, wherein in the step of determining that the user has installed the similarity of the application and the different information in the information stream library, the number of users who have clicked the information j among the users who have installed the application i and The number of users who install the application i determines the similarity between the user i installed the application i and the information j in the information flow library.
  19. 根据权利要求18所述的终端设备,其中在确定用户已安装应用与信息流库里的不同信息的相似度的步骤中,使用下列公式计算所述相似度:The terminal device according to claim 18, wherein in the step of determining that the user has installed the similarity of the application and the different information in the information stream library, the similarity is calculated using the following formula:
    Figure PCTCN2017119610-appb-100013
    Figure PCTCN2017119610-appb-100013
    其中:among them:
    s i,j表示已安装应用i与信息j的相似度,i=1,2,…,n j=1,2…,m, s i,j denotes the similarity between the installed application i and the information j, i=1, 2, . . . , n j=1, 2..., m,
    b i,j表示在已安装应用i的用户中点击过信息j的用户数量,i=1,2,…,n j=1,2…,m, b i,j represents the number of users who have clicked on the information j among the users who have installed the application i, i=1, 2, ..., n j=1, 2..., m,
    a i表示已安装应用i的用户数量,i=1,2,…,n, a i indicates the number of users who have installed the application i, i=1, 2,...,n,
    K i表示已安装应用i包含的标签的集合,i=1,2,…,n, K i indicates that the set of tags included in the application i has been installed, i=1, 2, ..., n,
    L j表示信息j包含的标签的集合,j=1,2,…,m; L j represents a set of labels contained in the information j, j = 1, 2, ..., m;
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    N为大于0的整数。N is an integer greater than zero.
  20. 根据权利要求17所述的终端设备,其中在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,根据用户最近一次安装应用i并保持到今天的天数和所获得的相似度来确定用户对信息j的匹配度,其中所获得的相似度为已安装应用i与信息流库里的信息j之间的相似度。The terminal device according to claim 17, wherein in the step of obtaining the degree of matching of the user with different information by using the obtained similarity, the number of days and the obtained similarity are maintained according to the last time the user installed the application i and maintained to today To determine the degree of matching of the user to the information j, wherein the similarity obtained is the similarity between the installed application i and the information j in the information stream library.
  21. 根据权利要求20所述的终端设备,其中在利用所获得的相似度来得到用户对不同信息的匹配度的步骤中,使用下列公式计算所述匹配度:The terminal device according to claim 20, wherein in the step of obtaining the degree of matching of the user with different information by using the obtained similarity, the matching degree is calculated using the following formula:
    Figure PCTCN2017119610-appb-100014
    Figure PCTCN2017119610-appb-100014
    其中:among them:
    u j表示用户对信息j的匹配度,j=1,2…,m, u j represents the degree of matching of the user to the information j, j=1, 2..., m,
    o i表示用户对应用i的兴趣度,i=1,2,…,n, o i indicates the user's interest in the application i, i=1, 2,...,n,
    s i,j表示已安装应用i与信息j的相似度,i=1,2,…,nj=1,2…,m, s i,j represents the similarity between the installed application i and the information j, i=1, 2,..., nj=1, 2...,m,
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述的用户对应用i的兴趣度o i的计算方法如下: The user's interest rate i i for the application i is calculated as follows:
    Figure PCTCN2017119610-appb-100015
    Figure PCTCN2017119610-appb-100015
    t i表示用户最近一次安装应用i并保持到今天的天数,i=1,2,…,n; t i represents the number of days the user last installed the application i and maintained to today, i=1, 2,...,n;
    n表示用户在到今天为止的N天内保持安装的应用的数量;n represents the number of applications the user has installed in the N days until today;
    N为大于0的整数。N is an integer greater than zero.
  22. 根据权利要求17所述的终端设备,其中还包括:根据预定的信息质量规则从候选信息里选取相应信息向用户推荐。The terminal device according to claim 17, further comprising: selecting corresponding information from the candidate information to recommend the user according to a predetermined information quality rule.
  23. 根据权利要求22所述的终端设备,其中在根据预定的信息质量规则从候选信息里选取相应信息向用户推荐的步骤中,根据候选信息的点击 率、点赞率和转化率之一、或者根据任意两两参数组合、或者根据该三个参数计算出的综合质量指标值,按综合质量指标值从大到小的顺序选取相应信息向用户推荐。The terminal device according to claim 22, wherein in the step of recommending the corresponding information from the candidate information to the user according to a predetermined information quality rule, according to one of a click rate, a click rate, and a conversion rate of the candidate information, or according to Any combination of two or two parameters, or the comprehensive quality index value calculated according to the three parameters, is selected according to the comprehensive quality index value from the largest to the smallest, and recommended to the user.
  24. 根据权利要求23所述的终端设备,其中所述综合质量指标值的计算方法为:The terminal device according to claim 23, wherein the calculation method of the comprehensive quality indicator value is:
    qul j=θ*ctr j+γ*gtr j+β*dtr j j=1,2,…,m Qul j =θ*ctr j +γ*gtr j +β*dtr j j=1,2,...,m
    其中:among them:
    qul j表示信息j的综合质量指标,j=1,2…,m, Qul j represents the comprehensive quality index of information j, j=1, 2..., m,
    ctr j表示信息j的点击率,j=1,2…,m, Ctr j represents the click rate of information j, j=1, 2..., m,
    gtr j表示信息j的点赞率,j=1,2…,m, Gtr j represents the praise rate of information j, j=1, 2..., m,
    dtr j表示信息j的转化率,j=1,2…,m, Dtr j represents the conversion rate of information j, j=1, 2..., m,
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    θ、γ和β是用来调节每个因素的权重,其中β+γ+θ=1,且β、γ和θ∈[0,1],通过取β、γ和θ不同值来确定通过那些参数指标来计算所述综合质量指标值;θ, γ, and β are used to adjust the weight of each factor, where β+γ+θ=1, and β, γ, and θ∈[0,1] are determined by taking different values of β, γ, and θ. a parameter indicator to calculate the comprehensive quality indicator value;
    所述信息j的点击率
    Figure PCTCN2017119610-appb-100016
    The click rate of the information j
    Figure PCTCN2017119610-appb-100016
    c j表示点击过信息j的用户数; c j represents the number of users who clicked on the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述信息j的点赞率
    Figure PCTCN2017119610-appb-100017
    The rate of praise of the information j
    Figure PCTCN2017119610-appb-100017
    g j表示点赞过信息j的用户数; g j represents the number of users who liked the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量;m represents the amount of information in the information flow library;
    所述信息j的转化率
    Figure PCTCN2017119610-appb-100018
    Conversion rate of the information j
    Figure PCTCN2017119610-appb-100018
    d j表示通过信息j产生下载应用行为的用户数; d j represents the number of users who generate the application behavior by the information j;
    f j表示向用户展示过信息j的所有用户数; f j represents the number of all users who have presented the information j to the user;
    m表示信息流库里的信息数量。m represents the amount of information in the stream library.
  25. 一种计算机可读存储介质,其上承载一个或多个计算机指令程序,所述计算机指令程序被一个或多个处理器执行时,所述一个或多个处理器执行权利要求1~8任一项所述的方法。A computer readable storage medium carrying one or more computer program programs, executed by one or more processors, the one or more processors executing any one of claims 1-8 The method described in the item.
PCT/CN2017/119610 2016-12-29 2017-12-28 Method and device for recommending application information based on installed application, terminal device, and storage medium WO2018121700A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201611242501.6 2016-12-29
CN201611242501.6A CN106846094A (en) 2016-12-29 2016-12-29 A kind of method and apparatus for recommending application message based on application has been installed

Publications (1)

Publication Number Publication Date
WO2018121700A1 true WO2018121700A1 (en) 2018-07-05

Family

ID=59113161

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/119610 WO2018121700A1 (en) 2016-12-29 2017-12-28 Method and device for recommending application information based on installed application, terminal device, and storage medium

Country Status (2)

Country Link
CN (1) CN106846094A (en)
WO (1) WO2018121700A1 (en)

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106846094A (en) * 2016-12-29 2017-06-13 广州优视网络科技有限公司 A kind of method and apparatus for recommending application message based on application has been installed
CN106897412A (en) * 2017-02-20 2017-06-27 广州优视网络科技有限公司 A kind of method and apparatus for recommending associated application based on intended application
CN110020152B (en) * 2017-12-04 2021-03-19 中国移动通信集团终端有限公司 Application recommendation method and device
CN108052671B (en) * 2017-12-29 2021-01-08 北京奇元科技有限公司 Method and system for pushing content items based on dynamic matching
CN108510352A (en) * 2018-02-09 2018-09-07 广州优视网络科技有限公司 Application issue recommends method, apparatus and computer equipment
CN108769125B (en) * 2018-04-28 2021-08-17 阿里巴巴(中国)有限公司 Application recommendation method and device, storage medium and computer equipment
CN108809987B (en) * 2018-06-14 2020-10-23 江苏果米文化发展有限公司 Online game popularization method based on big data analysis
CN109242643B (en) * 2018-09-30 2021-05-28 武汉斗鱼网络科技有限公司 Game recommendation method and device applied to live broadcast platform and storage medium
CN111796883B (en) * 2019-04-09 2024-05-14 Oppo广东移动通信有限公司 Equipment control method and device, storage medium and electronic equipment
CN110134410B (en) * 2019-05-17 2022-11-15 北京百度网讯科技有限公司 Personalized information association method and device for cold-start user and storage medium
CN111611143B (en) * 2020-05-20 2023-06-30 每日互动股份有限公司 Data processing method and device, electronic equipment and computer readable storage medium
CN113626694A (en) * 2020-10-29 2021-11-09 胡培培 Application program pushing system based on big data
CN115150346A (en) * 2022-07-04 2022-10-04 中国银行股份有限公司 Information pushing method and device

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103679502A (en) * 2013-11-14 2014-03-26 青岛海信传媒网络技术有限公司 Application recommendation method and system thereof
CN105912550A (en) * 2015-12-15 2016-08-31 乐视网信息技术(北京)股份有限公司 Method and device for information recommendation of mobile terminal
CN105989074A (en) * 2015-02-09 2016-10-05 北京字节跳动科技有限公司 Method and device for recommending cold start through mobile equipment information
CN106846094A (en) * 2016-12-29 2017-06-13 广州优视网络科技有限公司 A kind of method and apparatus for recommending application message based on application has been installed

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105989108A (en) * 2015-02-12 2016-10-05 广东欧珀移动通信有限公司 Information recommendation method and device
CN105426514B (en) * 2015-11-30 2019-03-01 扬州大学 Personalized mobile application APP recommended method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103679502A (en) * 2013-11-14 2014-03-26 青岛海信传媒网络技术有限公司 Application recommendation method and system thereof
CN105989074A (en) * 2015-02-09 2016-10-05 北京字节跳动科技有限公司 Method and device for recommending cold start through mobile equipment information
CN105912550A (en) * 2015-12-15 2016-08-31 乐视网信息技术(北京)股份有限公司 Method and device for information recommendation of mobile terminal
CN106846094A (en) * 2016-12-29 2017-06-13 广州优视网络科技有限公司 A kind of method and apparatus for recommending application message based on application has been installed

Also Published As

Publication number Publication date
CN106846094A (en) 2017-06-13

Similar Documents

Publication Publication Date Title
WO2018121700A1 (en) Method and device for recommending application information based on installed application, terminal device, and storage medium
US10693981B2 (en) Provisioning personalized content recommendations
US11263217B2 (en) Method of and system for determining user-specific proportions of content for recommendation
CN108829808B (en) Page personalized sorting method and device and electronic equipment
US9286391B1 (en) Clustering and recommending items based upon keyword analysis
US8893012B1 (en) Visual indicator based on relative rating of content item
CN112307344B (en) Object recommendation model, object recommendation method and device and electronic equipment
CN110929052A (en) Multimedia resource recommendation method and device, electronic equipment and storage medium
WO2015034850A2 (en) Feature selection for recommender systems
CN106407210B (en) A kind of methods of exhibiting and device of business object
US20160171589A1 (en) Personalized application recommendations
WO2015175384A1 (en) Query categorizer
US20160283952A1 (en) Ranking information providers
US20120215664A1 (en) Epurchase model
WO2018161710A1 (en) Method, device, and storage medium for recommending search term on basis of input search term
US20160196579A1 (en) Dynamic deep links based on user activity of a particular user
US20130103550A1 (en) Discovery of digital goods in an online marketplace
CN108351787B (en) Device-based filtering of content items associated with mobile applications
CN112136127A (en) Action indicator for search operation output element
US20210192549A1 (en) Generating analytics tools using a personalized market share
US10937070B2 (en) Collaborative filtering to generate recommendations
US20160125080A1 (en) Accessing Special Purpose Search Systems
JP5813052B2 (en) Information processing apparatus, method, and program
US9804741B2 (en) Methods and systems for managing N-streams of recommendations
US10242069B2 (en) Enhanced template curating

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17886584

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 13/08/2019)

122 Ep: pct application non-entry in european phase

Ref document number: 17886584

Country of ref document: EP

Kind code of ref document: A1