WO2018133680A1 - 一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质 - Google Patents

一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质 Download PDF

Info

Publication number
WO2018133680A1
WO2018133680A1 PCT/CN2018/071539 CN2018071539W WO2018133680A1 WO 2018133680 A1 WO2018133680 A1 WO 2018133680A1 CN 2018071539 W CN2018071539 W CN 2018071539W WO 2018133680 A1 WO2018133680 A1 WO 2018133680A1
Authority
WO
WIPO (PCT)
Prior art keywords
application
user
hot
applications
similarity
Prior art date
Application number
PCT/CN2018/071539
Other languages
English (en)
French (fr)
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 WO2018133680A1 publication Critical patent/WO2018133680A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation

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 computer readable storage medium for recommending hot words based on a user installed application.
  • search is the main entry point for users to download apps.
  • they need to recommend hot words to users in the search homepage.
  • Figure 1 shows an example of recommending hot words in the search homepage.
  • the hot words displayed on the search home page are often the hottest words currently searched, or hot words recommended based on the operation strategy.
  • the hot words recommended according to the prior art are not necessarily of interest to the user, and cannot meet the needs of different users, resulting in a poor user experience.
  • the embodiment of the present application provides a method for recommending a hot word based on a user installed application, which includes:
  • Determining a degree of matching of the user with the hot word based on a degree of matching of the user installed application with the hot word and a number of installed applications of the user;
  • a certain number of corresponding hot words are selected from the hot vocabulary as the recommended hot words.
  • the embodiment of the present application further provides an apparatus for recommending a hot word based on a user installed application, including:
  • the similarity determination unit of the application is used to determine the similarity between the installed application and the application in the application library
  • the application download probability determining unit is configured to determine a probability that the user downloads the application by searching for hot words in the hot word database;
  • a matching degree first determining unit configured to determine, according to the similarity and the probability, a matching degree between a user installed application and the hot word
  • a matching degree second determining unit determining a matching degree between the user and the hot word based on a matching degree of the user installed application and the hot word and a number of installed applications of the user;
  • the recommendation unit is configured to select a certain number of corresponding hot words from the hot vocabulary as the recommended hot words according to the matching degree values of the user and the hot words.
  • the formula for determining the similarity between the installed user and other applications includes:
  • n the number of applications in the application library
  • K i indicates that the user has installed the set of tags that the application i has
  • K j represents a set of tags that the application j in the application library has
  • U i represents a collection of users who have installed application i;
  • U j represents the set of users who installed the application j
  • the formula for determining the probability that the user downloads the application by searching for the hot words in the hot lexicon includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • a l,j represents the total number of users who downloaded the application j by searching for the hot word l;
  • w l represents the total number of users who searched for the hot word l.
  • the formula for determining, according to the similarity and the probability, that the user has installed the matching degree between the application and the hot word includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • P l,j represents the probability
  • the formula for determining the matching degree between the user and the hot word based on the matching degree between the user installed application and the hot word and the number of installed applications of the user includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • h u,i indicates whether the application i is installed in the application library, and the values are 1 and 0. The value of the application i is 1 and the value of the application i is 0.
  • the embodiment of the present application further provides a terminal device, including: a memory, storing a computer program; and a processor, executing the computer program, and implementing the following steps: determining a similarity between a user installed application and an application in the application library Determining a probability that the user downloaded the application by searching for hot words in the hot lexicon; determining a degree of matching between the user installed application and the hot word based on the similarity and the probability; based on the user having installed the application and The degree of matching between the hot words and the number of applications installed by the user to determine the degree of matching between the user and the hot words; and selecting a certain degree of matching value between the user and the hot words from the hot vocabulary The corresponding number of hot words is used as a recommended hot word.
  • the formula for determining the similarity between the installed user and other applications includes:
  • n the number of applications in the application library
  • K i indicates that the user has installed the set of tags that the application i has
  • K j represents a set of tags that the application j in the application library has
  • U i represents a collection of users who have installed application i;
  • U j represents the set of users who installed the application j
  • the formula for determining the probability that the user downloads the application by searching for the hot words in the hot lexicon includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • a l,j represents the total number of users who downloaded the application j by searching for the hot word l;
  • w l represents the total number of users who searched for the hot word l.
  • the formula for determining, according to the similarity and the probability, that the user has installed the matching degree between the application and the hot word includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • P l,j represents the probability
  • the formula for determining the matching degree between the user and the hot word based on the matching degree between the user installed application and the hot word and the number of installed applications of the user includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • h u,i indicates whether the application i is installed in the application library, and the values are 1 and 0. The value of the application i is 1 and the value of the application i is 0.
  • the embodiment of the present application provides a computer readable storage medium storing computer executable instructions for performing the above-described method of recommending a hot word based on a user installed application.
  • the method, device, terminal device and computer readable storage medium for recommending a hot word based on a user installed application by first determining the similarity between the user installed application and the application in the application library, and downloading the user through the hot word
  • the probability of the application is used to determine the degree of matching between the installed application and the hot word, and then the matching degree between the plurality of applications installed by the user and the related hot words is summarized into the matching degree between the user and the related hot words, thereby realizing the basis
  • the user's hobbies are personalized to recommend hot words, greatly improving the user experience.
  • FIG. 1 is a screenshot showing an example of recommending a hot word in a search home page according to the prior art
  • FIG. 2 is a screenshot showing an application with 2 tags displayed on the application market
  • FIG. 3 is a flowchart of a method for recommending a hot word based on a user installed application according to the first embodiment of the present application;
  • FIG. 4 is a schematic block diagram of an apparatus for recommending a hot word based on a user installed application according to a second embodiment of the present application.
  • Applicants of the technical solution fully consider the interests and hobbies of users, and propose a new personalized recommendation method, which can recommend different hot words according to different users' interests and hobbies, thereby realizing personalized recommendation, which will greatly enhance User experience.
  • FIG. 3 is a flow chart of a method for recommending a hot word based on a user installed application according to the first embodiment of the present application. As shown in FIG. 3, the method for recommending a hot word based on a user installed application of the present application includes the following steps:
  • the similarity between the installed application and all applications in the application library is determined according to the installed application on the terminal device used by the user.
  • the app library is a collection of all the apps available in the app marketplace or app store.
  • the installed application described here refers to an application that has been installed on a terminal used by the user when recommending a hot word to the user.
  • a very simple method such as classification the similarity of the application of the same type to the installed application is set to 1, and the similarity of the application of the different types is set to zero.
  • various applications referred to as applications
  • tags which are used to identify the classification or content of various applications, which is convenient for users to find.
  • each application in the application market or application store will contain at least one application tag, as shown in FIG. 2, and FIG. 2 shows that the entertainment application “everyday happy landlord” contains 2 tags, and 1 tag displays its logo.
  • the content of the application is "Double Landlord", and the other label shows that the classification that identifies the application is "card”.
  • the similarity of the application can be determined according to whether it has the same label as the installed application. Furthermore, the value of the similarity may be determined according to the number of the same specific tags. For example, the similarity of the application having one identical tag may be set to 1, and the similarity of the application having two identical tags is 2.
  • the above-exemplified method is the simplest method, and other methods can also be used.
  • a better determination method is provided, and the similarity value thus obtained is more representative of the similarity between the installed application and all applications in the application library.
  • the user may be determined based on a set of installation users of the application that the user has installed, a set of installation users of other applications, a label collection of an installation user set in which the user has installed the application, and a label collection of the installation user collection of other applications.
  • the similarity of installed apps to other apps is the simplest method, and other methods can also be used.
  • a better determination method is provided, and the similarity value thus obtained is more representative of the similarity between the installed application and all applications in the application library.
  • the user may be determined based on a set of installation users of the application that the user has installed, a set of installation users of other applications, a label collection of an installation user set in which the user has installed the application, and a label collection of the installation user collection of other applications.
  • the user is installed based on the installation user set of the user installed application, the installation user collection of other applications, the label collection of the installation user collection of the user installed application, and the label collection of the installation user collection of other applications.
  • the formula (1) that applies the similarity to the application in the application library includes:
  • n the number of applications in the application library
  • K i indicates that the user has installed the set of tags that the application i has
  • K j represents a set of tags that the application j in the application library has
  • U i represents a collection of users who have installed application i;
  • U j represents a collection of users who have installed application j
  • the above formula (1) considers two factors of similarity between two applications: 1 is the number factor with the same label, Value measurement, where Represents that each of the two applications has the same tag, the similarity increases by a factor of ⁇ , and the decrease of 1 is to classify the similarity between applications without the same tag as 0; 2 is the Jaked similarity coefficient, using Jaked formula
  • the Jakedian similarity coefficient is an indicator for measuring the similarity of two sets, that is, the similarity of the set of users who have installed the application i and the set of users who installed the application j.
  • S2 Determine the probability that the user downloads the application by searching for hot words in the hot lexicon.
  • the method for selecting a hot word to establish a hot vocabulary may be coexisting in multiple ways or in multiple ways, for example, one of the methods: statistic user inputting a search term, selecting a plurality of words whose first occurrence of the search term is ranked as a hot word.
  • Method 2 The words of the type to which the plurality of applications belong in the front are used as hot words; the third method: the top-ranking applications are summarized according to statistical data such as the user's rating or downloading behavior of the application, A number of words are summed up by the operating staff as hot words; in addition, any other method that can put a word into a hot vocabulary as a hot word can be used here.
  • the various applications installed on the smart mobile terminal used by the user may search for the application through hot words and download and install, or may not download and install the application through hot words, for example, the application is pre-installed by the smart mobile terminal; For example, the application is downloaded and installed directly by the user on the official website; and if the application is directly installed by the user through the installation package sent by the friend. Therefore, it is necessary to determine the probability that a user downloads an application by searching for hot words in the hot lexicon.
  • the probability P l,j can be calculated by equation (2):
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • a l,j represents the total number of users who downloaded the application j by searching for the hot word l;
  • w l represents the total number of users who searched for the hot word l.
  • the probability P l,j represents the proportion of users who have downloaded the application j through the hot word 1 among all users who searched for the hot word l in the hot lexicon.
  • S3 Determine, according to the similarity and the probability, a degree of matching between the user installed application and the hot word.
  • the formula (3) that determines the degree to which the user has installed the application and the hot word includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • P l,j represents the probability
  • S4 Determine a degree of matching between the user and the hot word based on the degree of matching between the user installed application and the hot word and the number of installed applications of the user.
  • the formula (4) for determining the degree of matching of the user with the hot word includes:
  • n the number of applications in the application library
  • n the number of hot words in the hot vocabulary
  • h u,i indicates whether the application i is installed in the application library, and the values are 1 and 0. The value of the application i is 1 and the value of the application i is 0.
  • S u,l represents the sum of the matching degrees of the plurality of applications installed by the user and the respective hot words.
  • S5 Select a certain number of corresponding hot words from the hot vocabulary as the recommended hot words according to the matching degree between the user and the hot words.
  • all the hot words in the hot vocabulary are arranged in descending order according to the matching degree S u, l from the largest to the smallest, starting from the hot word at the top.
  • a common application store or application market classifies hot words on the search home page, so a method of determining the category of each hot word l is also provided here, that is, by searching for the hot word l Among all the applications, the category with the most similar applications is set to the category of the hot word l.
  • the user has installed the application by first determining the similarity between the installed application and the application in the application library, and the probability that the user downloads the application through the hot word.
  • the matching degree of the hot words further summarizes the matching degree of the plurality of applications installed by the user and the related hot words into the matching degree between the user and the related hot words, thereby realizing the personalized recommendation hot words according to the user's interests and interests.
  • the purpose is to greatly improve the user experience.
  • FIG. 4 is a schematic block diagram of an apparatus for recommending a hot word based on a user installed application according to a second embodiment of the present application.
  • the apparatus for recommending a hot word based on a user installed application according to an embodiment of the present application includes:
  • the similarity determination unit of the application is used to determine the similarity between the installed application and the application in the application library
  • the application download probability determining unit is configured to determine a probability that the user downloads the application by searching for hot words in the hot word database;
  • a matching degree first determining unit configured to determine, according to the similarity and the probability, a matching degree between a user installed application and the hot word
  • a matching degree second determining unit determining a matching degree between the user and the hot word based on a matching degree of the user installed application and the hot word and a number of installed applications of the user;
  • the recommendation unit is configured to select a certain number of corresponding hot words from the hot vocabulary as the recommended hot words according to the matching degree values of the user and the hot words.
  • the application downloading probability determining unit For the specific working process of the included similarity determining unit, the application downloading probability determining unit, the matching degree first determining unit, the matching degree second determining unit, and the recommending unit, refer to the corresponding method steps S1-S5, where The description will not be repeated.
  • the device for recommending a hot word based on a user installed application further includes: a classification unit, configured to determine, in all applications downloaded by searching for the hot word l, a category with the highest number of similar applications as a hot word Category.
  • the user has installed the application by first determining the similarity between the installed application and the application in the application library, and the probability that the user downloads the application through the hot word.
  • the matching degree of the hot words further summarizes the matching degree of the plurality of applications installed by the user and the related hot words into the matching degree between the user and the related hot words, thereby realizing the personalized recommendation hot words according to the user's interests and interests.
  • the purpose is to greatly improve the user experience.
  • a computer program product for providing a method for recommending a hot word based on a user installed application which is a computer readable storage medium storing program code, the program code including instructions for performing the foregoing method implementation.
  • the program code including instructions for performing the foregoing method implementation.
  • the functions may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a standalone product.
  • the technical solution of the present invention which is essential or contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium, including The instructions are used to cause a computer device (which may be a personal computer, tablet, smartphone, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • a computer device which may be a personal computer, tablet, smartphone, server, or network device, etc.
  • the foregoing storage medium includes various media that can store program codes, such as a USB flash drive, a removable hard disk, a read only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.
  • the embodiment of the present application provides a computer readable storage medium storing computer executable instructions for performing the above-described method of recommending a hot word based on a user installed application.
  • the embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored on the memory and operable on the processor, where the processor is configured to implement the above-mentioned recommendation based on the user installed application when the program is executed.
  • the method of hot words including
  • the present application determines the degree of similarity between the installed application and the application in the application library, and the probability that the user downloads the application through the hot word to determine the matching degree between the installed application and the hot word, and then installs the user more.
  • the matching degree between the application and the related hot words is summarized into the matching degree between the user and the related hot words, thereby realizing the purpose of personalized recommendation hot words according to the user's interests and hobbies, and greatly improving the user experience.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质。所述方法包括:确定用户已安装应用与应用库里的应用的相似度(S1);确定用户通过搜索热词库里的热词而下载了应用的概率(S2);基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度(S3);基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度(S4);按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词(S5)。

Description

一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质
交互参考
本申请要求以下优先权:2017年01月17日提出的申请号:201710035726.2,名称:“一种基于用户已安装应用来推荐热词的方法和装置”的中国专利,本申请参考引用了如上所述申请的全部内容。
技术领域
本申请涉及信息处理技术领域,具体而言涉及一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质。
背景技术
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)。
在应用商店或者应用市场中,搜索是用户下载应用的主要入口。为了帮助用户发现更多有趣的应用,需要在搜索首页中对用户进行热词推荐,图1示出了在搜索首页中推荐显示热词的一个例子。现有技术中,搜索首页上展示的热词往往是目前搜索热度最高的词,或者基于运营策略而推荐的热词。但是,由于不同的用户拥有不同的兴趣,根据现有技术推荐的热词不一定是用户感兴趣的,无法满足不同用户的需求,致使用户的体验感不佳。
发明内容
本申请的目的在于提供一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质,以改善上述问题。
本申请实施例提供了一种基于用户已安装应用来推荐热词的方法,其包括:
确定用户已安装应用与应用库里的应用的相似度;
确定用户通过搜索热词库里的热词而下载了应用的概率;
基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;
基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及
按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
本申请实施例还提供了一种基于用户已安装应用来推荐热词的装置,其包括:
应用的相似度确定单元,用于确定用户已安装应用与应用库里的应用的相似度;
应用下载概率确定单元,用于确定用户通过搜索热词库里的热词而下载了应用的概率;
匹配度第一确定单元,用于基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;
匹配度第二确定单元,基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及
推荐单元,用于按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
其中,所述确定用户已安装应用与其他应用的相似度的公式包括:
Figure PCTCN2018071539-appb-000001
其中:
n表示应用库里的应用数量;
K i表示用户已安装应用i具有的标签集合;
K j表示应用库里的应用j具有的标签集合;
U i表示安装了应用i的用户集合;
U j表示安装了应用j的用户集合;以及
β>1。
其中,所述确定用户通过搜索热词库里的热词而下载了应用的概率的公式包括:
Figure PCTCN2018071539-appb-000002
其中:
n表示应用库里的应用数量;
m表示热词库里的热词数量;
a l,j表示通过搜索热词l而下载应用j的用户总数;以及
w l表示搜索了热词l的用户总数。
其中,所述基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度的公式包括:
Figure PCTCN2018071539-appb-000003
其中:
n表示表示应用库里的应用数量;
m表示表示热词库里的热词数量;
Sim i,j表示所述相似度;以及
P l,j表示所述概率。
其中,所述基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度的公式包括:
Figure PCTCN2018071539-appb-000004
其中:
n表示表示应用库里的应用数量;
m表示表示热词库里的热词数量;以及
h u,i表示用户安装应用库里的应用i与否,取值1和0,安装了应用i取值为1,没有安装应用i取值为0。
本申请实施例还提供了一种终端设备,其包括:存储器,储存计算机程序;以及处理器,执行所述计算机程序,并实现以下步骤:确定用户已安装应用与应用库里的应用的相似度;确定用户通过搜索热词库里的热词而下载了应用的概率;基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
其中,所述确定用户已安装应用与其他应用的相似度的公式包括:
Figure PCTCN2018071539-appb-000005
其中:
n表示应用库里的应用数量;
K i表示用户已安装应用i具有的标签集合;
K j表示应用库里的应用j具有的标签集合;
U i表示安装了应用i的用户集合;
U j表示安装了应用j的用户集合;以及
β>1。
其中,所述确定用户通过搜索热词库里的热词而下载了应用的概率的公式包括:
Figure PCTCN2018071539-appb-000006
其中:
n表示应用库里的应用数量;
m表示热词库里的热词数量;
a l,j表示通过搜索热词l而下载应用j的用户总数;以及
w l表示搜索了热词l的用户总数。
其中,所述基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度的公式包括:
Figure PCTCN2018071539-appb-000007
其中:
n表示表示应用库里的应用数量;
m表示表示热词库里的热词数量;
Sim i,j表示所述相似度;以及
P l,j表示所述概率。
其中,所述基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度的公式包括:
Figure PCTCN2018071539-appb-000008
其中:
n表示表示应用库里的应用数量;
m表示表示热词库里的热词数量;以及
h u,i表示用户安装应用库里的应用i与否,取值1和0,安装了应用i取值为1,没有安装应用i取值为0。
本申请实施例提供一种计算机可读存储介质,其存储有计算机可执行指令,所述计算机可执行指令用于执行上述的基于用户已安装应用来推荐热词的方法。
根据本申请的基于用户已安装应用来推荐热词的方法、装置、终端设备 及计算机可读存储介质,通过先确定用户已安装应用与应用库里的应用的相似度,和用户通过热词下载应用的概率,来确定用户已安装应用与所述热词的匹配度,再将用户安装的多个应用与相关热词的匹配度综合归纳为用户与相关热词的匹配度,从而实现了根据用户的兴趣爱好进行个性化推荐热词的目的,大大提高了用户体验。
附图说明
图1是示出根据现有技术在搜索首页中推荐热词的一个例子的截图;
图2是示例性的示出应用市场上显示的具有2个标签的应用的截图;
图3是本申请第一实施例的基于用户已安装应用来推荐热词的方法的流程图;以及
图4是本申请第二实施例的基于用户已安装应用来推荐热词的装置的示意性框图。
具体实施方式
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。
在上面提到的用户体验感不佳的原因之一是不同的用户拥有不同的兴趣,而现有技术的推荐方案仅仅是推荐搜索热度最高的词,但搜索热度最高的不一定是人人都喜欢的。以游戏为例,假设“捕鱼达人”是搜索热度最高的词,但用户A不喜欢玩“捕鱼达人”的游戏,而是喜欢玩格斗游戏,所以向用户A推荐“捕鱼达人”是无法激发他的兴趣去点击下载;再例如,用户B在朋友的推荐下下载了“捕鱼达人”,但不喜欢玩,将其卸载。但根据现有技术的 推荐方法,当用户B进入应用商店或者应用市场的搜索首页时依然会向其推荐“捕鱼达人”,这就带来了不好的用户体验。
本技术方案的申请人充分考虑了用户的兴趣和爱好,提出一种新的个性化推荐方法,能够根据不同用户的兴趣爱好不同而推荐的不同热词,从而实现个性化推荐,这会大大提升用户的体验感。
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智能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,如果能够找到一种方法可以基于用户已安装应用来推荐热词,就可以实现所述的个性化推荐的目的。
图3是本申请第一实施例的基于用户已安装应用来推荐热词的方法的流程图。如图3所示,本申请的基于用户已安装应用来推荐热词的方法包括以下步骤:
S1:确定用户已安装应用与应用库里的应用的相似度。
首先根据用户使用的终端设备上已安装的应用来确定已安装应用与应用库里的所有应用之间的相似度。应用库为应用市场或应用商店里提供的所有应用的集合。这里所述的已安装应用是指在向用户推荐热词时在用户使用的终端上已安装着的应用。
确定所述相似度的方法有很多,很简单的方法如分类法,将与该已安装应用同类的应用的相似度设为1,不同类的应用的相似度设为0。另外,应用商店或者应用市场里提供的各种应用程序(简称应用)通常都具有标签,标签的作用是标识各种应用程序的分类或内容,便于用户查找。目前,在应用市场或应用商店中每一个应用都会包含至少1个应用标签,如图2所示,图2示出了娱乐应用“天天欢乐斗地主”包含2个标签,1个标签显示其标识该应用的内容是“斗地主”,另1个标签显示其标识该应用的分类是“纸牌”。这样,可以根据是否与该已安装应用具有相同标签,来确定应用的相似度。再者,还可以根据具体相同标签的数量确定相似度的值,例如可以设具有1个相同标签的应用的相似度为1,设具有2个相同标签的应用的相似度为2。
当然,上述举例的方法为最简单的方法,也可以使用其他方法。在本实施例里提供一种更佳的确定方法,由此得到的相似度值更能表现已安装应用与应用库里的所有应用之间的相似度。具体地,本实施例中,可基于用户已安装应用的安装用户集合、其他应用的安装用户集合、用户已安装应用的安装用户集合的标签集合、其他应用的安装用户集合的标签集合,确定用户已安装应用与其他应用的相似度。
在一个优选的方案中,基于用户已安装应用的安装用户集合、其他应用的安装用户集合、用户已安装应用的安装用户集合的标签集合、其他应用的安装用户集合的标签集合,确定用户已安装应用与应用库里的应用的相似度的公式(1)包括:
Figure PCTCN2018071539-appb-000009
其中:
n表示应用库里的应用数量;
K i表示用户已安装应用i具有的标签集合;
K j表示应用库里的应用j具有的标签集合;
U i表示安装了应用i的用户集合;以及
U j表示安装了应用j的用户集合;
其中β>1,其取值考虑两个因素:1是应用库里的应用j具有的平均标签数量,设为k;2是标签相似权重,设为m,则
Figure PCTCN2018071539-appb-000010
根据经验,应用库里的应用j具有的平均标签数量k=2,标签相似权重m一般为4,因此β的经验值为2。
已知用户通常会在其使用的终端上安装多个应用,例如2个或更多。
上述公式(1)考虑了2个应用之间的相似度的两个因素:1是具有相同 标签的数量因素,用
Figure PCTCN2018071539-appb-000011
值衡量,其中
Figure PCTCN2018071539-appb-000012
表示2个应用之间每多一个相同标签,相似度增加β倍,而减1是为了把没有相同标签的应用之间的相似度归为0;2是杰卡德相似系数,用杰卡德公式
Figure PCTCN2018071539-appb-000013
衡量,杰卡德相似系数是衡量两个集合相似度的一种指标,即衡量已安装应用i的用户集合和安装应用j的用户集合的相似度。
这里简单举例说明具有相同标签的数量因素。例如,三个应用分别是应用A“斗地主”(其标签有“休闲”、“棋牌”、“斗地主”)、应用B“消消乐”(其标签有“休闲”“消除”)、应用C“德州扑克”(其标签有“休闲”“棋牌”);假设β=2,因为应用A和应用B之间具有1个相同标签,所以
Figure PCTCN2018071539-appb-000014
Figure PCTCN2018071539-appb-000015
因为应用A和应用C之间具有2个相同标签,所以
Figure PCTCN2018071539-appb-000016
这意味着在标签维度上应用A与应用C的相似度是应用A与应用B的相似度的3倍,即增加β=2倍。
当然,应理解,确定用户已安装应用与应用库里的应用的相似度的公式有多个,前面公式(1)仅展示出了一种优先的实现方式,在具体的应用中,还可能有其他的表现形式,例如,可以对公式(1)进行适当的变形,等等,本申请实施例对此不作限制。
S2:确定用户通过搜索热词库里的热词而下载了应用的概率。
在应用商店或者应用市场具有了推荐功能之后,开发商都会建立热词库,用于从热词库里选择出一定数量的热词向用户推荐。用于选择热词来建立热词库的方法可以是多种方式或多种方式并存,例如方法之一:统计用户输入搜索词,选择搜索词的出现次数排在前面的多个词作为热词;方法之二:将下载量排在前面的多个应用所属的类型的词作为热词;方法之三:根据用户对应用的评分或下载行为等统计数据对排名靠前的应用进行归纳整理,由运营工作人员归纳出多个词作为热词;除此之外,其他任何可以将某个词放入 热词库里作为热词的方法都可以在这里使用。
用户使用的智能移动终端上安装的各种应用可能是通过热词搜索到该应用并下载安装,也可能没有通过热词搜索该应用来下载安装,例如该应用是智能移动终端预安装的;再如该应用是用户直接在官网上下载安装的;再如也可以是用户通过朋友发送的安装包直接安装的。因此,有必要确定用户通过搜索热词库里的热词而下载了应用的概率。
可以通过公式(2)来计算所述概率P l,j
Figure PCTCN2018071539-appb-000017
其中:
n表示应用库里的应用数量;
m表示热词库里的热词数量;
a l,j表示通过搜索热词l而下载应用j的用户总数;以及
w l表示搜索了热词l的用户总数。
这样可知,概率P l,j表示在搜索了热词库里的热词l的所有用户中通过该热词l下载了应用j的用户所占比例。
S3:基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度。
在得到了用户已安装应用与应用库里的应用的相似度,和用户通过搜索热词库里的热词而下载了应用的概率之后,就可以确定用户已安装应用与所述热词的匹配度,即通过热词而下载了应用的所述热词。确定用户已安装应用与热词的匹配度的公式(3)包括:
Figure PCTCN2018071539-appb-000018
其中:
n表示表示应用库里的应用数量;
m表示表示热词库里的热词数量;
Sim i,j表示所述相似度;以及
P l,j表示所述概率。
S4:基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度。
因为通常情况下用户会在其使用的终端上安装超过1个数量的应用,例如安装有5个、10个或更多应用等,所以需要把已安装的每个应用与相关热词的匹配度综合归纳为用户与相关热词的匹配度。由此,确定用户与所述热词的匹配度的公式(4)包括:
Figure PCTCN2018071539-appb-000019
其中:
n表示表示应用库里的应用数量;
m表示表示热词库里的热词数量;以及
h u,i表示用户安装应用库里的应用i与否,取值1和0,安装了应用i取值为1,没有安装应用i取值为0。
这样可知,S u,l表示用户安装的多个应用与各自相应热词的匹配度累加之和。
S5:按用户与所述热词的匹配度从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
根据用户u与热词l的匹配度S u,l的大小,对热词库里的所有热词按匹配度S u,l从大到小做降序排列,从排在最前面的热词开始选取一定数量的热词作为推荐热词,展示给用户。所述一定数量在实践中可以自由选择,例如选择10个,或者20个或其他等。
在一个实施例中,常见的应用商店或者应用市场在搜索首页上都会给热词分类,因此这里也提供了一种确定每个热词l的类别的方法,即在通过搜索该热词l下载的所有应用中,同类应用数量最多的类别就设为该热词l的类别。
根据本申请的基于用户已安装应用来推荐热词的方法,通过先确定用户已安装应用与应用库里的应用的相似度,和用户通过热词下载应用的概率,来确定用户已安装应用与所述热词的匹配度,再将用户安装的多个应用与相关热词的匹配度综合归纳为用户与相关热词的匹配度,从而实现了根据用户的兴趣爱好进行个性化推荐热词的目的,大大提高了用户体验。
图4是本申请第二实施例的基于用户已安装应用来推荐热词的装置的示意性框图。如图4所示,本申请实施例的基于用户已安装应用来推荐热词的装置包括:
应用的相似度确定单元,用于确定用户已安装应用与应用库里的应用的相似度;
应用下载概率确定单元,用于确定用户通过搜索热词库里的热词而下载了应用的概率;
匹配度第一确定单元,用于基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;
匹配度第二确定单元,基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及
推荐单元,用于按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
其中,所包含的应用的相似度确定单元、应用下载概率确定单元、匹配度第一确定单元、匹配度第二确定单元和推荐单元的具体工作过程可以参见上述对应的方法步骤S1-S5,这里不再重复描述。
在一个实施例中,所述基于用户已安装应用来推荐热词的装置还包括:分类单元,用于在通过搜索热词l下载的所有应用中,确定同类应用数量最多 的类别为热词l的类别。
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,在此不再重复描述。
根据本申请的基于用户已安装应用来推荐热词的装置,通过先确定用户已安装应用与应用库里的应用的相似度,和用户通过热词下载应用的概率,来确定用户已安装应用与所述热词的匹配度,再将用户安装的多个应用与相关热词的匹配度综合归纳为用户与相关热词的匹配度,从而实现了根据用户的兴趣爱好进行个性化推荐热词的目的,大大提高了用户体验。
本申请实施例所提供的基于用户已安装应用来推荐热词的方法的计算机程序产品,其为包括存储了程序代码的计算机可读存储介质,所述程序代码包括的指令可用于执行前面方法实施例中所述的方法,具体实现可参见方法实施例,在此不再赘述。所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时,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,本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案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,平板电脑,智能手机,服务器,或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括:U盘、移动硬盘、只读存储器(ROM)、随机存取存储器(RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
本申请实施例提供一种计算机可读存储介质,其存储有计算机可执行指令,所述计算机可执行指令用于执行上述的基于用户已安装应用来推荐热词的方法。
本申请实施例提供一种终端设备,包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序,处理器配置为执行所述程序时实现上 述的基于用户已安装应用来推荐热词的方法。
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。
工业实用性
本申请通过先确定用户已安装应用与应用库里的应用的相似度,和用户通过热词下载应用的概率,来确定用户已安装应用与所述热词的匹配度,再将用户安装的多个应用与相关热词的匹配度综合归纳为用户与相关热词的匹配度,从而实现了根据用户的兴趣爱好进行个性化推荐热词的目的,大大提高了用户体验。

Claims (19)

  1. 一种基于用户已安装应用来推荐热词的方法,其包括:
    确定用户已安装应用与应用库里的应用的相似度;
    确定用户通过搜索热词库里的热词而下载了应用的概率;
    基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;
    基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及
    按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
  2. 根据权利要求1所述的方法,其中,所述确定用户已安装应用与其他应用的相似度包括:
    基于用户已安装应用的安装用户集合、其他应用的安装用户集合、用户已安装应用的安装用户集合的标签集合、其他应用的安装用户集合的标签集合,确定用户已安装应用与其他应用的相似度。
  3. 根据权利要求2所述的方法,其中,所述确定用户已安装应用与其他应用的相似度的公式包括:
    Figure PCTCN2018071539-appb-100001
    其中:
    n表示应用库里的应用数量;
    K i表示用户已安装应用i具有的标签集合;
    K j表示应用库里的应用j具有的标签集合;
    U i表示安装了应用i的用户集合;
    U :表示安装了应用j的用户集合;
    |K i∩K j|表示K i和标签集合K j之间相同标签的个数;
    Figure PCTCN2018071539-appb-100002
    表示用户集合U i和用户集合U j的杰卡德相似系数;
    Sim i,j表示用户已安装应用i和的应用库里的应用j相似度;
    以及
    β>1。
  4. 根据权利要求1所述的方法,其中,所述确定用户通过搜索热词库里的热词而下载了应用的概率的公式包括:
    Figure PCTCN2018071539-appb-100003
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    a l,j表示通过搜索热词l而下载应用j的用户总数;
    w l表示搜索了热词l的用户总数;以及
    P l,j表示通过搜索热词l而下载应用j的概率。
  5. 根据权利要求1所述的方法,其中,所述基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度的公式包括:
    Figure PCTCN2018071539-appb-100004
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    Sim i,j表示用户已安装应用i和应用库里的应用j的相似度;
    P l,j表示通过搜索热词l而下载应用j的概率;以及
    Wim i,l表示用户已安装应用i和搜索热词l的匹配度。
  6. 根据权利要求1所述的方法,其中,所述基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度的公式包括:
    Figure PCTCN2018071539-appb-100005
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    Wim i,l表示用户已安装应用i和搜索热词l的匹配度;
    h u,i表示用户安装应用库里的应用i与否,取值1和0,安装了应用i取值为1,没有安装应用i取值为0;以及
    S u,l表示用户u与搜索热词l的匹配度。
  7. 一种基于用户已安装应用来推荐热词的装置,其包括:
    应用的相似度确定单元,用于确定用户已安装应用与应用库里的应用的相似度;
    应用下载概率确定单元,用于确定用户通过搜索热词库里的热词而下载了应用的概率;
    匹配度第一确定单元,用于基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;
    匹配度第二确定单元,基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及
    推荐单元,用于按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
  8. 根据权利要求7所述的装置,其中,应用的相似度确定单元,具体用于:基于用户已安装应用的安装用户集合、其他应用的安装用户集合、用户已安装应用的安装用户集合的标签集合、其他应用的安装用户集合的标签集合,确定用户已安装应用与其他应用的相似度。
  9. 根据权利要求8所述的装置,其中,所述确定用户已安装应用与其他应用的相似度的公式包括:
    Figure PCTCN2018071539-appb-100006
    其中:
    n表示应用库里的应用数量;
    K i表示用户已安装应用i具有的标签集合;
    K j表示应用库里的应用j具有的标签集合;
    U i表示安装了应用i的用户集合;
    U j表示安装了应用j的用户集合;
    |K i∩K j|表示K i和标签集合K j之间相同标签的个数;
    Figure PCTCN2018071539-appb-100007
    表示用户集合U i和用户集合U j的杰卡德相似系数;
    Sim i,j表示用户已安装应用i和的应用库里的应用j相似度;以及
    β>1。
  10. 根据权利要求7所述的装置,其中,所述确定用户通过搜索热词库里的热词而下载了应用的概率的公式包括:
    Figure PCTCN2018071539-appb-100008
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    a l,j表示通过搜索热词l而下载应用j的用户总数;
    w l表示搜索了热词l的用户总数;以及
    P l,j表示通过搜索热词l而下载应用j的概率。
  11. 根据权利要求7所述的装置,其中,所述基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度的公式包括:
    Figure PCTCN2018071539-appb-100009
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    Sim i,j表示用户已安装应用i和应用库里的应用j的相似度;
    P l,j表示通过搜索热词l而下载应用j的概率;以及
    Wim i,l表示用户已安装应用i和搜索热词l的匹配度。
  12. 根据权利要求7所述的装置,其中,所述基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度的公式包括:
    Figure PCTCN2018071539-appb-100010
    其中:
    n表示表示应用库里的应用数量;
    m表示表示热词库里的热词数量;
    Wim i,l表示用户已安装应用i和搜索热词l的匹配度;
    h u,i表示用户安装应用库里的应用i与否,取值1和0,安装了应用i取值为1,没有安装应用i取值为0;以及
    S u,l表示用户u与搜索热词l的匹配度。
  13. 一种终端设备,其包括:
    存储器,储存计算机程序;以及
    处理器,执行所述计算机程序,并实现以下步骤:确定用户已安装应用与应用库里的应用的相似度;确定用户通过搜索热词库里的热词而下载了应用的概率;基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度;基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度;以及按用户与所述热词的匹配度值从大到小顺序从热词库里选取一定数量的相应热词作为推荐热词。
  14. 根据权利要求13所述的终端设备,其中,所述处理器具体用于:基于用户已安装应用的安装用户集合、其他应用的安装用户集合、用户已安装 应用的安装用户集合的标签集合、其他应用的安装用户集合的标签集合,确定用户已安装应用与其他应用的相似度。
  15. 根据权利要求14所述的终端设备,其中,所述确定用户已安装应用与其他应用的相似度的公式包括:
    Figure PCTCN2018071539-appb-100011
    其中:
    n表示应用库里的应用数量;
    K i表示用户已安装应用i具有的标签集合;
    K j表示应用库里的应用j具有的标签集合;
    U i表示安装了应用i的用户集合;
    U j表示安装了应用j的用户集合;
    |K i∩K j|表示K i和标签集合K j之间相同标签的个数;
    Figure PCTCN2018071539-appb-100012
    表示用户集合U i和用户集合U j的杰卡德相似系数;
    Sim i,j表示用户已安装应用i和的应用库里的应用j相似度;以及
    β>1。
  16. 根据权利要求13所述的终端设备,其中,所述确定用户通过搜索热词库里的热词而下载了应用的概率的公式包括:
    Figure PCTCN2018071539-appb-100013
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    a l,j表示通过搜索热词l而下载应用j的用户总数;
    w l表示搜索了热词l的用户总数;以及
    P l,j表示通过搜索热词l而下载应用j的概率。
  17. 根据权利要求13所述的终端设备,其中,所述基于所述相似度和所述概率确定用户已安装应用与所述热词的匹配度的公式包括:
    Figure PCTCN2018071539-appb-100014
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    Sim i,j表示用户已安装应用i和应用库里的应用j的相似度;
    P l,j表示通过搜索热词l而下载应用j的概率;以及
    Wim i,l表示用户已安装应用i和搜索热词l的匹配度。
  18. 根据权利要求13所述的终端设备,其中,所述基于所述用户已安装应用与所述热词的匹配度和用户已安装应用的数量来确定用户与所述热词的匹配度的公式包括:
    Figure PCTCN2018071539-appb-100015
    其中:
    n表示应用库里的应用数量;
    m表示热词库里的热词数量;
    Wim i,l表示用户已安装应用i和搜索热词l的匹配度;
    h u,i表示用户安装应用库里的应用i与否,取值1和0,安装了应用i取值为1,没有安装应用i取值为0;以及
    S u,l表示用户u与搜索热词l的匹配度。
  19. 一种计算机可读存储介质,其存储有计算机可执行指令,所述计算机可执行指令用于执行权利要求1至6中任一项所述的基于用户已安装应用来推荐热词的方法。
PCT/CN2018/071539 2017-01-17 2018-01-05 一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质 WO2018133680A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710035726.2 2017-01-17
CN201710035726.2A CN106844612B (zh) 2017-01-17 2017-01-17 一种基于用户已安装应用来推荐热词的方法和装置

Publications (1)

Publication Number Publication Date
WO2018133680A1 true WO2018133680A1 (zh) 2018-07-26

Family

ID=59123823

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/071539 WO2018133680A1 (zh) 2017-01-17 2018-01-05 一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质

Country Status (2)

Country Link
CN (1) CN106844612B (zh)
WO (1) WO2018133680A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106844612B (zh) * 2017-01-17 2021-04-09 阿里巴巴(中国)有限公司 一种基于用户已安装应用来推荐热词的方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103338223A (zh) * 2013-05-27 2013-10-02 清华大学 一种移动应用的推荐方法、客户端及服务器
CN103810030A (zh) * 2014-02-20 2014-05-21 北京奇虎科技有限公司 一种基于移动终端应用市场的应用推荐方法、装置及系统
CN104601659A (zh) * 2014-12-17 2015-05-06 深圳市腾讯计算机系统有限公司 应用推荐方法及系统
CN104808983A (zh) * 2015-03-19 2015-07-29 深圳市梦域科技有限公司 一种应用程序的推送方法及服务器
CN106844612A (zh) * 2017-01-17 2017-06-13 广州优视网络科技有限公司 一种基于用户已安装应用来推荐热词的方法和装置

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102622390A (zh) * 2011-10-11 2012-08-01 北京掌汇天下科技有限公司 一种移动终端中推荐应用的方法及服务器
CN104298679B (zh) * 2013-07-18 2019-05-07 腾讯科技(深圳)有限公司 应用业务推荐方法及装置
KR20150017156A (ko) * 2013-08-06 2015-02-16 삼성전자주식회사 휴대단말기에서 추천어를 제공하는 방법 및 장치
CN105069106A (zh) * 2015-08-07 2015-11-18 小米科技有限责任公司 应用群组推荐方法和装置
CN105893441A (zh) * 2015-12-15 2016-08-24 乐视网信息技术(北京)股份有限公司 一种终端的应用推荐方法及应用推荐系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103338223A (zh) * 2013-05-27 2013-10-02 清华大学 一种移动应用的推荐方法、客户端及服务器
CN103810030A (zh) * 2014-02-20 2014-05-21 北京奇虎科技有限公司 一种基于移动终端应用市场的应用推荐方法、装置及系统
CN104601659A (zh) * 2014-12-17 2015-05-06 深圳市腾讯计算机系统有限公司 应用推荐方法及系统
CN104808983A (zh) * 2015-03-19 2015-07-29 深圳市梦域科技有限公司 一种应用程序的推送方法及服务器
CN106844612A (zh) * 2017-01-17 2017-06-13 广州优视网络科技有限公司 一种基于用户已安装应用来推荐热词的方法和装置

Also Published As

Publication number Publication date
CN106844612B (zh) 2021-04-09
CN106844612A (zh) 2017-06-13

Similar Documents

Publication Publication Date Title
JP6615362B2 (ja) ユーザー似顔絵を取得する方法及び装置
WO2018121700A1 (zh) 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
US10332184B2 (en) Personalized application recommendations
WO2017101389A1 (zh) 一种移动终端的信息推荐方法及装置
US20170308950A1 (en) Systems and methods for programatically classifying text using topic classification
US11836778B2 (en) Product and content association
WO2020155877A1 (zh) 信息推荐
WO2019061989A1 (zh) 贷款风险控制方法、电子装置及可读存储介质
WO2018161710A1 (zh) 一种基于输入搜索词来推荐搜索词的方法、装置和存储介质
WO2015034850A2 (en) Feature selection for recommender systems
CN108304426B (zh) 标识的获取方法及装置
US9542480B2 (en) Systems and methods for programatically classifying text using category filtration
CN106452809B (zh) 一种数据处理方法和装置
WO2018171288A1 (zh) 给信息流标注标签的方法、装置、终端设备及存储介质
WO2018149244A1 (zh) 一种基于目标应用推荐相关联应用的方法和装置
US9201968B2 (en) System and method for finding mood-dependent top selling/rated lists
TW201926087A (zh) 問題推送方法及裝置
WO2018133680A1 (zh) 一种基于用户已安装应用来推荐热词的方法、装置、终端设备及计算机可读存储介质
US20180046683A1 (en) Search word list providing device and method using same
US11062371B1 (en) Determine product relevance
CN110020118B (zh) 一种计算用户之间相似度的方法及装置
CN111767457A (zh) 推荐方法及装置
JP6680472B2 (ja) 情報処理装置、情報処理方法及び情報処理プログラム
US20170068901A1 (en) Recommendation method and apparatus
JP5613699B2 (ja) 商品推薦装置及び方法及びプログラム

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: 18741221

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 1205A DATED 24.10.19.

122 Ep: pct application non-entry in european phase

Ref document number: 18741221

Country of ref document: EP

Kind code of ref document: A1