WO2018149244A1 - 一种基于目标应用推荐相关联应用的方法和装置 - Google Patents
一种基于目标应用推荐相关联应用的方法和装置 Download PDFInfo
- Publication number
- WO2018149244A1 WO2018149244A1 PCT/CN2017/120183 CN2017120183W WO2018149244A1 WO 2018149244 A1 WO2018149244 A1 WO 2018149244A1 CN 2017120183 W CN2017120183 W CN 2017120183W WO 2018149244 A1 WO2018149244 A1 WO 2018149244A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- application
- target application
- target
- user
- click
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/90335—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/9032—Query formulation
- G06F16/90324—Query formulation using system suggestions
Definitions
- the present invention relates to the field of information processing technologies, and in particular, to a method and apparatus for recommending an associated application based on a target application.
- a common recommendation presentation method is to recommend an associated application according to the application that the user currently clicks, for example, "Everyone also downloads" shown in FIG. 1A, and "Downloaded *** will also download” as shown in FIG. 1B.
- the recommended logic for such scenarios is to recommend a batch of associated applications to the user based on the current application.
- the traditional recommendation method is to adopt the label collaborative filtering method, that is, firstly, the recommended application has the same label as the target application, and then the user behavior space vector of each application is established through the user's download, browse, and installed behavior data, and finally The cosine coefficient (or Jacques coefficient, Pearson coefficient, etc.) calculates the similarity value between the recommended application and the target application, and takes the top application of the similarity ranking as the recommended candidate application.
- the above recommendation method has a disadvantage in the application related application recommendation scenario: the main idea is that which application is downloaded by the person who downloads the target application, and the focus of the consideration is to find out which applications are more relevant from the behavior of the user.
- the recommended application may be “Golden and Jade, and it is ruined”. The reason is that some poor quality applications do a good job on the packaging, and many users are attracted by it. Click behavior, which causes the existing recommendation method to think that the application is very popular with users and recommend it, but in fact the experience of these applications is very poor.
- the above-mentioned existing recommendation method has a drawback in that it is easy to recommend an application with a poor experience to the user, resulting in a user's feeling of experience in the application store or the application market installed by the user.
- An embodiment of the present invention provides a method for recommending an associated application based on a target application, including:
- the searched associated applications are ranked in descending order based on the matching degree and are sequentially recommended to the user.
- the application having the same tag as the tag of the target application is searched for.
- the at least one parameter comprises determining the matching degree according to a parameter of any one or any combination of the heat, the quality of the score, the click rate and the conversion rate of the associated application.
- An embodiment of the present invention further provides an apparatus for recommending an associated application based on a target application, including:
- a search unit for searching for an application associated with the target application according to the tag
- a matching degree determining unit configured to select an appropriate one or more parameters to determine a matching degree of the searched related application for the target application
- a recommendation unit configured to sort the searched related applications in descending order based on the matching degree and sequentially recommend to the user.
- the search unit is configured to search for an application having the same tag as the tag of the target application.
- the one or more parameters include: one of the heat, the score quality, the click rate, and the conversion rate of the associated application relative to the target application, or any combination of 2 parameters, or any combination of 3 parameters, or The four parameter combinations, or other parameters.
- hot(j) represents the popularity of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Install(i,j) indicates whether the user i of the target application has installed the associated application j, and the value is 0 or 1, 0 means that the associated application j is not installed, and 1 means that the associated application j is installed;
- evl(j) represents the quality of the rating of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Evaluate (i, j) represents the evaluation score of the user i of the installed target application for the associated application j, which is an integer between -1 or [0, 5], and when the value is -1, the user is not given Out of the evaluation score.
- ctr(j) represents the click rate of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Click(i,j) indicates whether the user i who has installed the target application generates a click behavior on the associated application j displayed to it, and click(i,j) takes the value -1, 0, 1, where -1 indicates no user i shows the associated application j, 0 means that the associated application j is displayed to the user i but the user i does not click the associated application j, 1 indicates that the associated application j is displayed to the user i and the user i clicks on the associated application j;
- dtr(j) represents the conversion rate of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Down(i,j) indicates whether the user i who has installed the target application generates a download behavior for the associated application j displayed to it.
- the value of down(i,j) is -1, 0, 1, where -1 indicates no user i shows the associated application j, 0 means to show the associated application j to the user i but the user i did not download the associated application j, 1 means to show the associated application j to the user i and the user i downloaded the associated application j;
- the method for determining the matching degree of the searched related application for the target application is as follows:
- fit(j) represents the degree of matching of the associated application j with respect to the target application
- Avg(hot(j)) represents the average of the heats of all associated applications retrieved
- Avg(evl(j)) represents the average of the score quality of all associated applications retrieved
- Avg(ctr(j)) represents the average of the click rates of all associated applications retrieved
- Avg(dtr(j)) represents the average of the conversion rates of all associated applications retrieved
- an embodiment of the present invention further provides a computing device, including:
- a memory for storing information
- the searched associated applications are ranked in descending order based on the matching degree and are sequentially recommended to the user.
- the formulation is also formulated for:
- the formulation is also formulated for:
- the at least one parameter includes determining the degree of matching according to a parameter of any one or any combination of the heat, the score quality, the click rate, and the conversion rate of the associated application with respect to the target application.
- the formulation is also formulated for:
- hot(j) represents the heat of the associated application j relative to the target application
- m represents the number of associated applications retrieved
- n represents the number of users who have installed the target application
- install(i,j) indicates installed Whether the user i of the target application has the associated application j installed.
- the formulation is also formulated for:
- evl(j) represents the quality of the rating of the associated application j relative to the target application
- m represents the number of associated applications retrieved
- n represents the number of users who have installed the target application
- evaluate(i,j) indicates installed The evaluation score of the user i of the target application for the associated application j.
- the formulation is also formulated for:
- ctr(j) represents the click rate of the associated application j relative to the target application
- m represents the number of associated applications retrieved
- n represents the number of users who have installed the target application
- click(i,j) indicates installed Whether the user i of the target application generates a click behavior for the associated application j that is presented to it.
- the formulation is also formulated for:
- dtr(j) represents the conversion rate of the associated application j relative to the target application
- m represents the number of associated applications retrieved
- n represents the number of users who have installed the target application
- down(i,j) indicates installed Whether the user i of the target application generates a download behavior for the associated application j that is presented to it.
- the feature of the matching degree is described as:
- fit(j) represents the degree of matching of the associated application j with respect to the target application
- avg(hot(j)) represents the average of the heat of the associated application
- avg(evl(j)) represents the quality of the rating of the associated application.
- Average of the values; avg(ctr(j)) represents the average of the click-through rates of the associated applications, avg(dtr(j)) represents the average of the conversion rates of the associated applications, and ⁇ , ⁇ , ⁇ , and ⁇ represent the weighting parameters .
- an embodiment of the present invention further provides a computer readable storage medium carrying one or more computer instruction programs, where the computer instruction program is executed by one or more processors, the one or A plurality of processors perform the method of recommending an associated application based on the target application as described above.
- an application associated with a target application is first retrieved through a tag, and the associated method is that the retrieved application has the same label as the target application, and then comprehensively considers the search.
- the relevance of the associated application relative to the target application's popularity, click-through rate, conversion rate, rating quality, or other parameters to measure the matching degree between the associated application and the target application and finally recommend the associated application with a large matching degree based on the matching degree.
- the application recommended by this scheme not only considers the similarity factor, but also considers the quality of the recommended application, and improves the deficiencies of the prior art which is easy to recommend the application with poor experience to the user, and improves the user experience.
- FIGS. 1A and 1B are schematic screenshots showing a recommended associated application according to a recommendation method of the prior art
- FIG. 2 is a schematic flow chart of a method for recommending an associated application based on a target application of the present invention
- FIG. 3 is a screenshot showing an application with 2 tags displayed on the application market.
- FIG. 4 is a schematic block diagram of an apparatus for recommending an associated application based on a target application of the present invention.
- Figure 5 is a block diagram showing the basic structure of a terminal device of the present invention.
- FIG. 2 is a schematic flow diagram of a method of recommending an associated application based on a target application. As shown in FIG. 2, the method for recommending an associated application based on a target application of the present invention includes the following steps:
- S1 Search for an application associated with the target application based on the tag.
- each application in the application market or application store will contain at least one application tag, as shown in FIG. 3, and FIG. 3 shows that the entertainment application "Everyday 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”.
- an application having the same tag as the tag of the target application may be searched as an associated application, and the target application refers to an application currently clicked by the user.
- the related applications retrieved according to the present invention include The application of any of the same tags, that is, the retrieved associated applications, includes all applications that have the tag content as "landlord" and all applications that have the tag content "card”.
- the application with the same label as the target application is retrieved, so that the recommended application is first limited to the related application of the same category, so that the recommended application has a certain degree of similarity with the target application, and the subsequent calculation amount can be reduced.
- S2 Select an appropriate one or more parameters to determine the degree of matching of the searched associated application for the target application.
- the application is recommended only in a similar manner, and it is easy to recommend an application with a poor experience. This requires consideration of the quality of the associated application retrieved in the previous step.
- the matching degree of the searched associated application for the target application may be calculated by some parameter indicators, and the quality of the associated application is measured according to the value of the matching degree.
- the appropriate one or more parameters described herein include: one of the heat, the score quality, the click rate, and the conversion rate of the associated application relative to the target application, or any two parameter combinations thereof, or any three parameters Combination, or combination of the four parameters, or other parameters.
- the parameter when one parameter is selected, the parameter may be the heat, the score quality, the click rate, or the conversion rate of the associated application relative to the target application, or other parameters.
- the heat is the probability that a user who installed the target application installs an associated application
- the rating quality is an average rating score of a user who has installed the target application for an associated application
- the click rate is the ratio of the number of times a user who installed the target application clicks on an associated application to the number of times it is displayed;
- the conversion rate is the ratio of the number of times a user who installed the target application downloads a related application to the number of times it is displayed.
- hot(j) represents the popularity of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Install(i,j) indicates whether the user i who has installed the target application has the associated application j installed, and the value is 0 or 1, 0 means that the associated application j is not installed, and 1 means that the associated application j is installed.
- the heat that can be derived from this formula is the probability that a user who installed the target application installs an associated application j.
- evl(j) represents the quality of the rating of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Evaluate (i, j) represents the evaluation score of the user i of the installed target application for the associated application j, which is an integer between -1 or [0, 5], and when the value is -1, the user is not given Out of the evaluation score.
- the quality of the score that can be derived from the formula is the average evaluation score of the user who installed the target application for an associated application j, that is, among all the users who installed the target application, regardless of whether an associated application j is not considered.
- the user data of the score is evaluated, and the average score of the remaining users for an associated application j is obtained.
- ctr(j) represents the click rate of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Click(i,j) indicates whether the user i who has installed the target application generates a click behavior on the associated application j displayed to it, and click(i,j) takes the value -1, 0, 1, where -1 indicates no user i shows the associated application j, 0 means that the associated application j is displayed to the user i but the user i does not click the associated application j, 1 indicates that the associated application j is displayed to the user i and the user i clicks on the associated application j;
- the click rate that can be derived from the formula is the ratio of the number of times the user who installed the target application clicks on an associated application to the number of times it is displayed, that is, among all the users who installed the target application, regardless of the user i has not been shown. Correlate the user data of the application j, and the resulting click rate for a certain associated application j.
- dtr(j) represents the conversion rate of the associated application j relative to the target application
- n the number of associated applications retrieved
- n indicates the number of users who have installed the target application
- Down(i,j) indicates whether the user i who has installed the target application generates a download behavior for the associated application j displayed to it.
- the value of down(i,j) is -1, 0, 1, where -1 indicates no user i shows the associated application j, 0 means to show the associated application j to the user i but the user i did not download the associated application j, 1 means to show the associated application j to the user i and the user i downloaded the associated application j;
- the conversion rate that can be derived from this formula is the ratio of the number of times a user who installed the target application downloads a related application to the number of times it is displayed, that is, among all the users who installed the target application, regardless of the fact that the user i has not been shown. Applying the user data of j, the resulting conversion rate for a certain associated application j.
- the method of selecting the appropriate one or more parameters to determine the matching degree of the searched associated application for the target application is as follows:
- fit(j) represents the degree of matching of the associated application j with respect to the target application
- Avg(hot(j)) represents the average of the heats of all associated applications retrieved
- Avg(evl(j)) represents the average of the score quality of all associated applications retrieved
- Avg(ctr(j)) represents the average of the click rates of all associated applications retrieved
- Avg(dtr(j)) represents the average of the conversion rates of all associated applications retrieved
- the purpose of dividing the value of each factor (heat, score quality, click rate, and conversion rate) by the average of the factors is to normalize the factor value.
- one of the heat, the score quality, the click rate and the conversion rate of the associated application relative to the target application can be selected, or Any combination of 2 parameters, or any combination of 3 parameters, or a combination of the 4 parameters, determines the degree of matching of the searched associated application for the target application.
- S3 Sort the searched related applications in descending order based on the matching degree and sequentially recommend to the user.
- the searched application associated with the target application is sorted in descending order according to the matching degree value, and then the result is used as the associated recommended application list of the target application, and displayed in the order of the list in the page.
- an application associated with a target application is first retrieved through a tag, and the associated method is that the retrieved application has the same label as the target application, and then comprehensively considers the retrieved association.
- the application s relativeness to the target app’s popularity, clickthrough rate, conversion rate, rating quality, or other parameters, where the heat factor is considered from the user’s perspective, and the click-through rate factor is considered to be attractive to the user, using conversions.
- the quality of the rate and the score is to consider the quality of the application, so as to measure the matching degree between the related application and the target application, and finally recommend the related application with large matching degree according to the size of the matching degree.
- the application recommended by this scheme not only considers the similarity factor, but also considers the quality of the recommended application, and improves the deficiencies of the prior art which is easy to recommend the application with poor experience to the user, and improves the user experience.
- the apparatus for recommending an associated application based on a target application of the present invention includes:
- a search unit for searching for an application associated with the target application according to the tag
- a matching degree determining unit configured to select an appropriate one or more parameters to determine a matching degree of the searched related application for the target application
- a recommendation unit configured to sort the searched related applications in descending order based on the matching degree and sequentially recommend to the user.
- the search unit searching for the application associated with the target application according to the label, refer to the corresponding method step, that is, the search unit preferably searches for an application having the same label as the target application's label.
- the specific implementation process of the matching degree determining unit for selecting the appropriate one or more parameters to determine the matching degree of the searched related application for the target application may also be referred to the corresponding method step.
- the corresponding method step The specific implementation process of the matching degree determining unit for selecting the appropriate one or more parameters to determine the matching degree of the searched related application for the target application.
- the appropriate one or more parameters described herein include: one of the heat, the score quality, the click rate, and the conversion rate of the associated application relative to the target application, or any two parameter combinations thereof, or any three parameters Combination, or combination of the four parameters, or other parameters.
- the heat is the probability that a user who installed the target application installs an associated application
- the rating quality is an average rating score of a user who has installed the target application for an associated application
- the click rate is the ratio of the number of times a user who installed the target application clicks on an associated application to the number of times it is displayed;
- the conversion rate is the ratio of the number of times a user who installed the target application downloads a related application to the number of times it is displayed.
- the calculation method of matching degree is also the same as described above, that is,
- fit(j) represents the degree of matching of the associated application j with respect to the target application
- Avg(hot(j)) represents the average of the heats of all associated applications retrieved
- Avg(evl(j)) represents the average of the score quality of all associated applications retrieved
- Avg(ctr(j)) represents the average of the click rates of all associated applications retrieved
- Avg(dtr(j)) represents the average of the conversion rates of all associated applications retrieved
- the different values of ⁇ and ⁇ are used to determine, by those parameter indicators, the degree of matching of the searched associated application to the target application.
- the purpose of dividing the value of each factor (heat, score quality, click rate, and conversion rate) by the average of the factors is to normalize the factor value.
- the recommendation unit ranks the searched related applications in descending order based on the matching degree and sequentially recommends to the user.
- the application associated with the target application is first retrieved through the label, and the associated method is that the retrieved application has the same label as the target application, and then comprehensively considers the retrieved association.
- the application s relativeness to the target app’s popularity, clickthrough rate, conversion rate, rating quality, or other parameters, where the heat factor is considered from the user’s perspective, and the click-through rate factor is considered to be attractive to the user, using conversions.
- the quality of the rate and the score is to consider the quality of the application, so as to measure the matching degree between the related application and the target application, and finally recommend the related application with large matching degree according to the size of the matching degree.
- the application recommended by this scheme not only considers the similarity factor, but also considers the quality of the recommended application, and improves the deficiencies of the prior art which is easy to recommend the application with poor experience to the user, and improves the user experience.
- a computer program product for a method for recommending an associated application based on a target application provided by an embodiment of the present invention comprising a computer readable storage medium storing program code, the program code comprising instructions operable to perform the foregoing method embodiment
- program code comprising instructions operable to perform the foregoing method embodiment
- FIG. 5 is a block diagram showing the basic structure of the terminal device according to the embodiment.
- 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 operation of the terminal device, and the processor 310 is configured to perform the method of recommending an associated application based on the target application provided by the above embodiments.
- the processor 310 is configured to:
- the searched associated applications are ranked in descending order based on the matching degree and are sequentially recommended to the user.
- the memory 320 is a non-volatile storage medium, and stores an operating system 321, a database 322, and a computer program for implementing the read-and-write separation mode-based download speed-up method provided by the above embodiments, and a candidate intermediate generated by executing the computer program. Data and result data.
- Network interface 340 is used to communicate with the server, and network interface 340 includes a radio frequency transceiver.
- the application further provides a computer readable storage medium having one or more computer program programs thereon, the one or more computer program programs being executed by one or more processors, the one or more The processor executes the method of recommending an associated application based on the target application as described above.
- 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.
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Stored Programmes (AREA)
Abstract
一种基于目标应用推荐相关联应用的方法和装置。所述方法包括:根据标签来搜索与目标应用相关联的应用(S1);选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度(S2);基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐(S3)。
Description
本发明涉及信息处理技术领域,具体而言涉及一种基于目标应用推荐相关联应用的方法和装置。
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)。
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。一种常见的推荐展示方式是根据用户当前点击的应用推荐相关联应用,例如图1A所示的“大家还下载”、图1B所示的“下载了***的人还会下载”。这类场景的推荐逻辑是根据当前应用给用户推荐一批相关联的应用。传统的推荐方法是采用标签协同过滤方法,即:首先限定推荐的应用在与目标应用有相同标签,然后通过用户的下载、浏览、已安装等行为数据建立各应用的用户行为空间向量,最后根据余弦系数(或杰卡德系数、皮尔森系数等)计算推荐应用与目标应用的相似度值,取相似度排名最前面的一批应用作为推荐候选应用。
但上述的推荐方法在应用关联应用推荐场景中存在不足在于:其主要思想是下载了目标应用的人还会下载哪些应用,考虑的重点是从用户的行为出发去发现哪些应用相关性更高,但缺乏对应用质量本身的考虑,这样推荐的应用可能是“金玉其外,败絮其中”,其原因是:有些质量不好的应用在包装上做得很好,很多用户因此被它吸引而产生点击行为,这会造成现有推荐方法认为这个应用很受用户欢迎而把它推荐出来,而实质上这些应用的体验感很差。简而言之,上述的现有推荐方法的缺点在于容易把体验感很差的应用推荐给用户,造成用户对其安装的应用商店或应用市场的体验感变差。
发明内容
本发明的目的在于一种基于目标应用推荐相关联应用的方法和装置,以改善上述问题。
本发明实施例提供了一种基于目标应用推荐相关联应用的方法,其包括:
根据标签来搜索与目标应用相关联的应用;
选择适当的至少一个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;
基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
优选的,在根据标签来搜索与目标应用相关联的应用的步骤中,搜索出具有与目标应用的标签相同标签的应用。
优选地,所述至少一个参数包括:根据所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率中的任意一种或任意多种结合的参数确定所述匹配度。
本发明实施例还提供了一种基于目标应用推荐相关联应用的装置,其包括:
搜索单元,用于根据标签来搜索与目标应用相关联的应用;
匹配度确定单元,用于选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;
推荐单元,用于基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
优选的,所述搜索单元用于搜索出具有与目标应用的标签相同标签的应用。
其中,所述1个或多个参数包括:所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率之一、或任意2个参数组合、或任意3个参数组合、或该4个参数组合、或者其他参数。
其中,1)所述相关联应用相对于目标应用的热度计算如下:
其中hot(j)表示相关联应用j相对于目标应用的热度;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
install(i,j)表示已安装目标应用的用户i是否安装了关联应用j,取值0或1,0表示未安装关联应用j,1表示安装了关联应用j;
2)所述相关联应用相对于目标应用的评分质量计算如下:
其中evl(j)表示相关联应用j相对于目标应用的评分质量;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
evaluate(i,j)表示已安装目标应用的用户i对关联应用j的评价分数,其取值为-1或[0,5]之间的整数,当取值为-1时表示用户没有给出评价分数。
函数if()在变量evaluate(i,j)=-1时为0,否则if()为1;
函数if'()在变量evaluate(i,j)=-1时为0,否则if'()为evaluate(i,j);
3)所述相关联应用相对于目标应用的点击率计算如下:
其中ctr(j)表示相关联应用j相对于目标应用的点击率;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
click(i,j)表示已安装目标应用的用户i是否对向其展示的关联应用j产生点击行为,click(i,j)取值为-1、0、1,其中-1表示没有向用户i展示过关联应用j,0表示向用户i展示了关联应用j但用户i没有点击该关联应用j,1表示向用户i展示关联应用j且用户i点击了该关联应用j;
函数if()在变量click(i,j)=-1时为0,否则if()为1;
函数if'()在变量click(i,j)=-1时为0,否则if'()为click(i,j);
4)所述相关联应用相对于目标应用的转化率计算如下:
其中dtr(j)表示相关联应用j相对于目标应用的转化率;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
down(i,j)表示已安装目标应用的用户i是否对向其展示的关联应用j产生下载行为,down(i,j)取值为-1、0、1,其中-1表示没有向用户i展示过关联应用j,0表示向用户i展示了关联应用j但用户i没有下载该关联应用j,1表示向用户i展示关联应用j且用户i下载了该关联应用j;
函数if()在变量down(i,j)=-1时为0,否则if()为1;
函数if'()在变量down(i,j)=-1时为0,否则if'()为down(i,j)。
其中,确定所搜索的相关联应用对于所述目标应用的匹配度的方法如下:
其中:fit(j)表示相关联应用j相对于目标应用的匹配度;
avg(hot(j))表示检索出来的所有相关联应用的热度的平均值;
avg(evl(j))表示检索出来的所有相关联应用的评分质量的平均值;
avg(ctr(j))表示检索出来的所有相关联应用的点击率的平均值;
avg(dtr(j))表示检索出来的所有相关联应用的转化率的平均值;
α、β、γ和θ是用来调节每个因素的权重,其中α+β+γ+θ=1,且α、β、γ和θ∈[0,1],通过取α、β、γ和θ不同值来确定通过那些参数指标来计算所搜索的相关联应用对于所述目标应用的匹配度。
为解决上述技术问题,本发明实施例还提供一种计算设备,包括:
存储器,用于存储信息;以及
连接至所述存储器的处理器,用于:
根据标签来搜索与目标应用相关联的应用;
选择适当的至少一个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;
基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
优选地,所述还被配制用于:
搜索出具有与目标应用的标签相同标签的应用。
优选地,所述还被配制用于:
所述至少一个参数包括:根据所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率中的任意一种或任意多种结合的参数确定所述匹配度。
优选地,所述还被配制用于:
所述目标应用的热度特征描述为:
其中,其中hot(j)表示相关联应用j相对于目标应用的热度,m表示检索出来的相关联应用的数量,n表示已安装目标应用的用户的数量,install(i,j)表示已安装目标应用的用户i是否安装了关联应用j。
优选地,所述还被配制用于:
所述目标应用的评分质量特征描述为:
其中,evl(j)表示相关联应用j相对于目标应用的评分质量,m表示检索出来的相关联应用的数量,n表示已安装目标应用的用户的数量,evaluate(i,j)表示已安装目标应用的用户i对关联应用j的评价分数。
优选地,所述还被配制用于:
所述目标应用的点击率特征描述为:
其中,ctr(j)表示相关联应用j相对于目标应用的点击率,m表示检索出来的相关联应用的数量,n表示已安装目标应用的用户的数量,click(i,j)表示已安装目标应用的用户i是否对向其展示的关联应用j产生点击行为。
优选地,所述还被配制用于:
所述目标应用的转化率特征描述为:
其中,dtr(j)表示相关联应用j相对于目标应用的转化率,m表示检索出来的相关联应用的数量,n表示已安装目标应用的用户的数量,down(i,j)表示已安装目标应用的用户i是否对向其展示的关联应用j产生下载行为。
优选地,所述匹配度的特征描述为:
其中,fit(j)表示相关联应用j相对于目标应用的匹配度,avg(hot(j))表示相关联应用的热度的平均值,avg(evl(j))表示相关联应用的评分质量的平均值;avg(ctr(j))表示相关联应用的点击率的平均值,avg(dtr(j))表示相关联应用的转化率的平均值,α、β、γ和θ表示权重参数。
为解决上述技术问题,本发明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执行上述所述的基于目标应用推荐相关联应用的方法。
根据本发明的基于目标应用推荐相关联应用的方法和装置,首先通过标签检索出与目标应用相关联的应用,其关联方式是检索出来的应用与目标应用具有相同的标签,然后综合考虑检索出来的关联应用相对于目标应用的热度、点击率、转化率、评分质量或其他参数等因素,用以衡量关联应用与目标应用的匹配度,最后根据匹配度的大小优先推荐匹配度大的关联应用。以这种方案推荐出来的应用,既考虑了相似度因素,又考虑了推荐应用的质量,改善了现有技术容易把体验感很差的应用推荐给用户的不足,提高了用户体验。
图1A和1B是根据现有技术的推荐方法展示推荐相关联应用的示意性截图;
图2是本发明的基于目标应用推荐相关联应用的方法的示意性流程图;
图3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4是本发明的基于目标应用推荐相关联应用的装置的示意性框图。
图5本发明的终端设备的基本结构框图。
下面将结合本发明实施例和附图,对本发明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发明一部分实施例,而 不是全部的实施例。通常在此处附图中描述和示出的本发明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发明的实施例的详细描述并非旨在限制要求保护的本发明的范围,而是仅仅表示本发明的选定实施例。基于本发明的实施例,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,都属于本发明保护的范围。
下面通过实施例详细描述本发明所提供的基于目标应用推荐相关联应用的方法和装置。
图2是一种基于目标应用推荐相关联应用的方法的示意性流程图。如图2所示,本发明的基于目标应用推荐相关联应用的方法包括以下步骤:
S1:根据标签来搜索与目标应用相关联的应用。
通常,应用商店或者应用市场里提供的各种应用程序(简称应用)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3所示,图3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”。
因此在本步骤中,在根据标签来搜索与目标应用相关联的应用的实现方法里可以搜索出具有与目标应用的标签相同标签的应用作为相关联应用,所谓目标应用就是指用户当前点击的应用。以图3所示为例,既可以搜索出标签内容为“斗地主”的所有应用,也可以是搜索出标签内容为“纸牌”的所有应用,而依据本发明检索出来的相关联应用包括包含任意相同标签的应用,即检索出来的相关联应用包括具有标签内容为“斗地主”的所有应用,和标签内容为“纸牌”的所有应用这两部分。
首先检索出与目标应用有相同标签的应用,从而先将推荐应用限定为同类别的相关联应用,可以使推荐应用与目标应用具有一定的相似度,还可以减少后面的运算量。
S2:选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度。
正如在上面阐述的,仅仅以相似度的方式推荐应用,容易把体验感很差的应用推荐出来。这就需要考虑上一步检索出来的相关联应用的质量。可以通过一些参数指标来计算所搜索的相关联应用对于所述目标应用的匹配度,根据匹配度的值来衡量相关联应用的质量。
这里所述适当的1个或多个参数包括:所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率之一、或其中的任意2个参数组合、或任意3个参数组合、或该4个参数组合、或者其他参数。
具体而言,当选择1个参数时,该参数可以是所述相关联应用相对于目标应用的热度、评分质量、点击率或者转化率,或者其他参数。
其中,所述热度就是安装了目标应用的用户安装某一个相关联应用的概率;
所述评分质量就是安装了目标应用的用户对某一个相关联应用的平均评价分数;
所述点击率就是安装了目标应用的用户点击某一个相关联应用的次数与其被显示次数之比;
所述转化率就是安装了目标应用的用户下载某一个相关应用的次数与其被显示次数之比。
1)所述相关联应用相对于目标应用的热度计算如下:
其中hot(j)表示相关联应用j相对于目标应用的热度;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
install(i,j)表示已安装目标应用的用户i是否安装了关联应用j,取值0或1,0表示未安装关联应用j,1表示安装了关联应用j。
由该公式可以得出的热度就是安装了目标应用的用户安装某一个相关 联应用j的概率。
在下面计算评分质量、点击率和转化率时会使用函数if()和if'(),其中
该函数if()的运算如下:
该函数if'()的运算如下:
2)所述相关联应用相对于目标应用的评分质量计算如下:
其中evl(j)表示相关联应用j相对于目标应用的评分质量;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
evaluate(i,j)表示已安装目标应用的用户i对关联应用j的评价分数,其取值为-1或[0,5]之间的整数,当取值为-1时表示用户没有给出评价分数。
函数if(evaluate(i,j))在evaluate(i,j)=-1时为0,否则if(evaluate(i,j))为1;
函数if'(evaluate(i,j))在evaluate(i,j)=-1时为0,否则if'(evaluate(i,j))为evaluate(i,j)。
由该公式可以得出的评分质量就是安装了目标应用的用户对某一个相关联应用j的平均评价分数,即在安装了目标应用的所有用户中,不考虑没有对某一相关联应用j进行评价分数的用户数据,所得出的其余用户对某一个相关联应用j的平均评价分数。
3)所述相关联应用相对于目标应用的点击率计算如下:
其中ctr(j)表示相关联应用j相对于目标应用的点击率;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
click(i,j)表示已安装目标应用的用户i是否对向其展示的关联应用j产生点击行为,click(i,j)取值为-1、0、1,其中-1表示没有向用户i展示过关联应用j,0表示向用户i展示了关联应用j但用户i没有点击该关联应用j,1表示向用户i展示关联应用j且用户i点击了该关联应用j;
函数jf(click(i,j))在click(i,j)=-1时为0,否则if(click(i,j))为1;
函数if'(click(i,j))在click(i,j)=-1时为0,否则if'(click(i,j))为click(i,j)。
由该公式可以得出的点击率就是安装了目标应用的用户点击某一个相关联应用的次数与其被显示次数之比,即在安装了目标应用的所有用户中,不考虑没有向用户i展示过关联应用j的用户数据,所得出的对某一个相关联应用j的点击率。
4)所述相关联应用相对于目标应用的转化率计算如下:
其中dtr(j)表示相关联应用j相对于目标应用的转化率;
m表示检索出来的相关联应用的数量;
n表示已安装目标应用的用户的数量;
down(i,j)表示已安装目标应用的用户i是否对向其展示的关联应用j产生下载行为,down(i,j)取值为-1、0、1,其中-1表示没有向用户i展示过关联应用j,0表示向用户i展示了关联应用j但用户i没有下载该关联应用j,1表示向用户i展示关联应用j且用户i下载了该关联应用j;
函数if(down(i,j))在down(i,j)=-1时为0,否则if(down(i,j))为1;
函数if'(down(i,j))在down(i,j)=-1时为0,否则if'(down(i,j))为down(i,j)。
由该公式可以得出的转化率就是安装了目标应用的用户下载某一个相关应用的次数与其被显示次数之比,即在安装了目标应用的所有用户中,不考虑没有向用户i展示过关联应用j的用户数据,所得出的对某一个相关联应用j的转化率。
这样,选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度的方法如下:
其中:fit(j)表示相关联应用j相对于目标应用的匹配度;
avg(hot(j))表示检索出来的所有相关联应用的热度的平均值;
avg(evl(j))表示检索出来的所有相关联应用的评分质量的平均值;
avg(ctr(j))表示检索出来的所有相关联应用的点击率的平均值;
avg(dtr(j))表示检索出来的所有相关联应用的转化率的平均值;
将每个因素(热度、评分质量、点击率和转化率)的值除以该因素的平均值的目的是对该因素值做标准化处理。
α、β、γ和θ是用来调节每个因素的权重,其中α+β+γ+θ=1,且α、β、γ和θ∈[0,1],可以通过取α、β、γ和θ不同值来确定通过那些参数指标来计算所搜索的相关联应用对于所述目标应用的匹配度。例如,当α=1、β=0、γ=0和θ=0时,则通过热度指标来计算匹配度,当α=0、β=1、γ=0和θ=0时,则通过评分质量指标来计算匹配度,以此类推,当α=0、β=0、γ=0.5和θ=0.5(γ和θ也可以取非0的其它值,且满足γ+θ=1即可)时,则通过点击率和转化率指标来计算匹配度,也可以让α=0.5、β=0.5、γ=0和θ=0(α和β也可以取非0的其它值,且满足α+β=1即可)或者α=0、β=0.5、γ=0.5和θ=0(β和γ也可以取非0的其它值,且满足β+γ=1即可),还可以选取其他组合;再如, 可以让α=0.5、β=0.3、γ=0.2和θ=0(α、β和γ也可以取非0的其它值,且满足α+β+γ=1即可),此时通过热度、评分质量和点击率指标来计算匹配度,也可以宣州区其他组合;如果将这4种因素同时考虑,则α+β+γ+θ=1,α、β、γ和θ∈(0,1],即α、β、γ和θ均不为0;如果这4种因素是同等重要,可以取α=β=γ=θ=0.25。由这些例子也可以得出,可以选择所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率之一、或其中的任意2个参数组合、或任意3个参数组合、或该4个参数组合,来确定所搜索的相关联应用对于所述目标应用的匹配度。
S3:基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
对所搜索的与目标应用相关联的应用根据匹配度值从大到小做降序排列,然后把结果作为目标应用的关联推荐应用列表,在页面中按列表顺序依次展示。
根据本发明的基于目标应用推荐相关联应用的方法,首先通过标签检索出与目标应用相关联的应用,其关联方式是检索出来的应用与目标应用具有相同的标签,然后综合考虑检索出来的关联应用相对于目标应用的热度、点击率、转化率、评分质量或其他参数等因素,其中使用热度因素是从用户的角度考虑相似性,使用点击率因素是考虑应用对用户的吸引力,使用转化率和评分质量是考虑应用的质量,从而衡量关联应用与目标应用的匹配度,最后根据匹配度的大小优先推荐匹配度大的关联应用。以这种方案推荐出来的应用,既考虑了相似度因素,又考虑了推荐应用的质量,改善了现有技术容易把体验感很差的应用推荐给用户的不足,提高了用户体验。
图4是本发明的基于目标应用推荐相关联应用的装置的示意性框图。 如图4所示,本发明的基于目标应用推荐相关联应用的装置包括:
搜索单元,用于根据标签来搜索与目标应用相关联的应用;
匹配度确定单元,用于选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;
推荐单元,用于基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
其中,搜索单元根据标签来搜索与目标应用相关联的应用的具体实现过程可以参见上述对应的方法步骤,即所述搜索单元优选搜索出具有与目标应用的标签相同标签的应用。
匹配度确定单元选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度的具体实现过程也可以参见上述对应的方法步骤。这里再简述一下。
这里所述适当的1个或多个参数包括:所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率之一、或其中的任意2个参数组合、或任意3个参数组合、或该4个参数组合、或者其他参数。
其中,所述热度就是安装了目标应用的用户安装某一个相关联应用的概率;
所述评分质量就是安装了目标应用的用户对某一个相关联应用的平均评价分数;
所述点击率就是安装了目标应用的用户点击某一个相关联应用的次数与其被显示次数之比;
所述转化率就是安装了目标应用的用户下载某一个相关应用的次数与其被显示次数之比。
所述的热度、评分质量、点击率和转化率的计算方法参见上面介绍的详细过程。这里不再赘述。
匹配度的计算方法也与上面介绍的相同,即
其中:fit(j)表示相关联应用j相对于目标应用的匹配度;
avg(hot(j))表示检索出来的所有相关联应用的热度的平均值;
avg(evl(j))表示检索出来的所有相关联应用的评分质量的平均值;
avg(ctr(j))表示检索出来的所有相关联应用的点击率的平均值;
avg(dtr(j))表示检索出来的所有相关联应用的转化率的平均值;
α、β、γ和θ是用来调节每个因素的权重,其中α+β+γ+θ=1,且α、β、γ和θ∈[0,1],可以通过取α、β、γ和θ不同值来确定通过那些参数指标来计算所搜索的相关联应用对于所述目标应用的匹配度。
将每个因素(热度、评分质量、点击率和转化率)的值除以该因素的平均值的目的是对该因素值做标准化处理。
推荐单元基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
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。
根据本发明的基于目标应用推荐相关联应用的装置,首先通过标签检索出与目标应用相关联的应用,其关联方式是检索出来的应用与目标应用具有相同的标签,然后综合考虑检索出来的关联应用相对于目标应用的热度、点击率、转化率、评分质量或其他参数等因素,其中使用热度因素是从用户的角度考虑相似性,使用点击率因素是考虑应用对用户的吸引力,使用转化率和评分质量是考虑应用的质量,从而衡量关联应用与目标应用的匹配度,最后根据匹配度的大小优先推荐匹配度大的关联应用。以这种方案推荐出来的应用,既考虑了相似度因素,又考虑了推荐应用的质量,改善了现有技术容易把体验感很差的应用推荐给用户的不足,提高了用户体验。
本发明实施例所提供的基于目标应用推荐相关联应用的方法的计算机 程序产品,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。
请参阅图5,图5为本实施例终端设备的基本结构框图。
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目标应用推荐相关联应用的方法。处理器310用于:
根据标签来搜索与目标应用相关联的应用;
选择适当的至少一个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;
基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。。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读写分离模式的下载提速方法的计算机程序,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。
本申请还提供一种计算机可读存储介质,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上述所述的基于目标应用推荐相关联应用的方法。
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。
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。
需要指出的是本实施例中的内存存储器中存储有用于执行安卓系统下适配应用通知颜色的方法的所有程序,处理器能够调用该内存储器中的程序实现上述安卓系统下适配应用通知颜色的方法的所有功能,由于安卓系统下适配应用通知颜色的方法在上述文件中已经进行详述,在此不再赘述。
以上所述,仅为本发明的具体实施方式,但本发明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发明揭露的技术范围内,可轻易想到变化或替换,都应涵盖在本发明的保护范围之内。因此,本发明的保护范围应以所述权利要求的保护范围为准。
Claims (25)
- 一种基于目标应用推荐相关联应用的方法,其包括:根据标签来搜索与目标应用相关联的应用;选择适当的至少一个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
- 根据权利要求1所述的方法,在根据标签来搜索与目标应用相关联的应用的步骤中包括:搜索出具有与目标应用的标签相同标签的应用。
- 根据权利要求1所述的方法,所述至少一个参数包括:根据所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率中的任意一种或任意多种结合的参数确定所述匹配度。
- 一种基于目标应用推荐相关联应用的装置,其包括:搜索单元,用于根据标签来搜索与目标应用相关联的应用;匹配度确定单元,用于选择适当的1个或多个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;推荐单元,用于基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
- 根据权利要求9所述的装置,所述搜索单元用于搜索出具有与目标应用的标签相同标签的应用。
- 根据权利要求9所述的装置,所述至少一个参数包括:根据所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率中的任意一种或任意多种结合的参数确定所述匹配度。
- 一种计算设备,包括:存储器,用于存储信息;以及连接至所述存储器的处理器,用于:根据标签来搜索与目标应用相关联的应用;选择适当的至少一个参数来确定所搜索的相关联应用对于所述目标应用的匹配度;基于所述匹配度大小对所搜索的相关联应用降序排列并且顺序向用户推荐。
- 根据权利要求17所述的计算设备,所述还被配制用于:搜索出具有与目标应用的标签相同标签的应用。
- 根据权利要求17所述的计算设备,所述还被配制用于:所述至少一个参数包括:根据所述相关联应用相对于目标应用的热度、评分质量、点击率和转化率中的任意一种或任意多种结合的参数确定所述匹配度。
- 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任一项所述的基于目标应用推荐相关联应用的方法。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710090023.XA CN106897412A (zh) | 2017-02-20 | 2017-02-20 | 一种基于目标应用推荐相关联应用的方法和装置 |
CN201710090023.X | 2017-02-20 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2018149244A1 true WO2018149244A1 (zh) | 2018-08-23 |
Family
ID=59185920
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2017/120183 WO2018149244A1 (zh) | 2017-02-20 | 2017-12-29 | 一种基于目标应用推荐相关联应用的方法和装置 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN106897412A (zh) |
WO (1) | WO2018149244A1 (zh) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106897412A (zh) * | 2017-02-20 | 2017-06-27 | 广州优视网络科技有限公司 | 一种基于目标应用推荐相关联应用的方法和装置 |
CN108279946B (zh) * | 2017-12-01 | 2022-01-25 | 运鼎科技(北京)有限公司 | 一种调取销售者应用的方法和装置 |
CN110163460B (zh) * | 2018-03-30 | 2023-09-19 | 腾讯科技(深圳)有限公司 | 一种确定应用分值的方法及设备 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103886090A (zh) * | 2014-03-31 | 2014-06-25 | 北京搜狗科技发展有限公司 | 基于用户喜好的内容推荐方法及装置 |
CN105893440A (zh) * | 2015-12-15 | 2016-08-24 | 乐视网信息技术(北京)股份有限公司 | 一种关联应用的推荐方法及装置 |
CN105938494A (zh) * | 2016-04-14 | 2016-09-14 | 乐视控股(北京)有限公司 | 一种应用搜索方法和装置 |
CN106846094A (zh) * | 2016-12-29 | 2017-06-13 | 广州优视网络科技有限公司 | 一种基于已安装应用来推荐应用信息的方法和装置 |
CN106897412A (zh) * | 2017-02-20 | 2017-06-27 | 广州优视网络科技有限公司 | 一种基于目标应用推荐相关联应用的方法和装置 |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102591876A (zh) * | 2011-01-14 | 2012-07-18 | 阿里巴巴集团控股有限公司 | 搜索结果排序方法及装置 |
CN104158865A (zh) * | 2014-08-01 | 2014-11-19 | 北京奇虎科技有限公司 | 一种推送应用到终端的方法和系统 |
CN105160003B (zh) * | 2015-09-11 | 2019-06-25 | 中山大学 | 一种基于地理位置的app检索排序方法及系统 |
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 |
-
2017
- 2017-02-20 CN CN201710090023.XA patent/CN106897412A/zh active Pending
- 2017-12-29 WO PCT/CN2017/120183 patent/WO2018149244A1/zh active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103886090A (zh) * | 2014-03-31 | 2014-06-25 | 北京搜狗科技发展有限公司 | 基于用户喜好的内容推荐方法及装置 |
CN105893440A (zh) * | 2015-12-15 | 2016-08-24 | 乐视网信息技术(北京)股份有限公司 | 一种关联应用的推荐方法及装置 |
CN105938494A (zh) * | 2016-04-14 | 2016-09-14 | 乐视控股(北京)有限公司 | 一种应用搜索方法和装置 |
CN106846094A (zh) * | 2016-12-29 | 2017-06-13 | 广州优视网络科技有限公司 | 一种基于已安装应用来推荐应用信息的方法和装置 |
CN106897412A (zh) * | 2017-02-20 | 2017-06-27 | 广州优视网络科技有限公司 | 一种基于目标应用推荐相关联应用的方法和装置 |
Also Published As
Publication number | Publication date |
---|---|
CN106897412A (zh) | 2017-06-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
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 | |
US9208437B2 (en) | Personalized information pushing method and device | |
KR102102568B1 (ko) | 최적화된 항목 속성 비교 | |
US8166062B1 (en) | Search-caching and threshold alerting for commerce sites | |
US9043351B1 (en) | Determining search query specificity | |
US20170293695A1 (en) | Optimizing similar item recommendations in a semi-structured environment | |
US20140114901A1 (en) | System and method for recommending application resources | |
WO2015034850A2 (en) | Feature selection for recommender systems | |
TWI662495B (zh) | 推廣資訊的處理方法、裝置及系統 | |
US20190294259A1 (en) | Sticker recommendation method and apparatus | |
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 | |
WO2020155877A1 (zh) | 信息推荐 | |
US20180012284A1 (en) | Information analysis apparatus, information analysis method, and non-transitory computer readable storage medium | |
US20130103550A1 (en) | Discovery of digital goods in an online marketplace | |
US9542482B1 (en) | Providing items of interest | |
WO2018149244A1 (zh) | 一种基于目标应用推荐相关联应用的方法和装置 | |
CN107423308A (zh) | 主题推荐方法以及装置 | |
US9785712B1 (en) | Multi-index search engines | |
US8666914B1 (en) | Ranking non-product documents | |
US9607100B1 (en) | Providing inline search suggestions for search strings | |
US20240354353A1 (en) | Method and apparatus for displaying search result, and method and apparatus for processing search request | |
CN114820123A (zh) | 团购商品推荐方法、装置、设备及存储介质 | |
WO2018171288A1 (zh) | 给信息流标注标签的方法、装置、终端设备及存储介质 | |
US9959559B2 (en) | Ranking and recommendation algorithms for search listings | |
WO2016149051A1 (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: 17897026 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 17897026 Country of ref document: EP Kind code of ref document: A1 |
|
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 05.12.2019. |