WO2017118426A1 - 社交平台的用户影响力估算方法、装置及计算机存储介质 - Google Patents

社交平台的用户影响力估算方法、装置及计算机存储介质 Download PDF

Info

Publication number
WO2017118426A1
WO2017118426A1 PCT/CN2017/070503 CN2017070503W WO2017118426A1 WO 2017118426 A1 WO2017118426 A1 WO 2017118426A1 CN 2017070503 W CN2017070503 W CN 2017070503W WO 2017118426 A1 WO2017118426 A1 WO 2017118426A1
Authority
WO
WIPO (PCT)
Prior art keywords
influence
user
ranking
social platform
estimating
Prior art date
Application number
PCT/CN2017/070503
Other languages
English (en)
French (fr)
Inventor
谭奔
刘大鹏
曹孝卿
张小鹏
肖磊
Original Assignee
腾讯科技(深圳)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 腾讯科技(深圳)有限公司 filed Critical 腾讯科技(深圳)有限公司
Publication of WO2017118426A1 publication Critical patent/WO2017118426A1/zh
Priority to US15/933,891 priority Critical patent/US20180211335A1/en

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/01Social networking
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241Advertisements
    • G06Q30/0251Targeted advertisements
    • G06Q30/0269Targeted advertisements based on user profile or attribute
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L51/00User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
    • H04L51/52User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail for supporting social networking services
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/2866Architectures; Arrangements
    • H04L67/30Profiles
    • H04L67/306User profiles
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/535Tracking the activity of the user

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to a method and apparatus for estimating user influence of a social platform.
  • the determination of user influence is generally based on the network of contacts.
  • users can add their favorite people as friends and even as close friends. Therefore, the calculation method based on the influence of the network is to calculate the coverage of the user's friends. The more friends a user has, the more friends there are. The higher its social influence.
  • User influence describes the ability of a user to influence other users. In the social network field (such as WeChat friends circle), user influence can be measured by the degree of attention received by the user. The higher the degree of attention, the social influence. It is bigger.
  • the existing user influence estimation scheme can estimate the social influence of the user to a certain extent, if the number of friends of the user is large, but the contact is rare, the social of the user is measured based on the coverage of the friend alone. Influence, will make the calculated user social Impact accuracy and credibility are not high, resulting in inaccurate information delivery on social platforms.
  • An object of the present invention is to provide a method and device for estimating user influence of a social platform, which aims to improve the accuracy and credibility of a user's social influence calculation, thereby improving the accuracy of information delivery on a social platform.
  • the embodiment of the present invention provides the following technical solutions:
  • a method for estimating user influence of a social platform including:
  • the influence of each user is determined according to the influence ranking.
  • the embodiment of the present invention further provides the following technical solutions:
  • a user influence estimation device for a social platform comprising:
  • An obtaining unit configured to acquire user behavior data on a social platform
  • a first determining unit configured to determine, according to the user behavior data, an influence transfer relationship between the two of the users
  • An estimating unit configured to estimate a ranking of influence of the user on the social platform based on the influence transfer relationship
  • the second determining unit is configured to determine the influence of each user according to the influence ranking.
  • a computer storage medium comprising a set of instructions that, when executed, cause at least one processor to perform operations comprising:
  • the influence of each user is determined according to the influence ranking.
  • the embodiment of the present invention first determines the influence transfer relationship between the two users according to the user behavior data on the social platform, and then estimates the influence of the user on the social platform based on the influence transfer relationship.
  • Ranking so that the influence of the user can be determined according to the influence ranking; since the user behavior data mainly reflects the interaction information of the user in the social activity, the scheme mainly determines the influence transfer relationship between the users based on the user behavior data, and is based on The influence transfer relationship estimates the influence of the user, so the accuracy and credibility of the user's social influence estimation are greatly improved compared to the existing method of measuring the social influence of the user based on the degree of friend coverage. , which also improves the accuracy of information delivery on social platforms.
  • 1a is a schematic diagram of a scenario of a method for estimating user influence of a social platform according to an embodiment of the present invention
  • 1b is a schematic flowchart of a method for estimating user influence of a social platform according to an embodiment of the present invention
  • 2a is a schematic flowchart of a method for estimating user influence of a social platform according to an embodiment of the present invention
  • 2b is a schematic diagram of application of a method for estimating user influence of a social platform according to an embodiment of the present invention
  • FIG. 3 is a schematic structural diagram of a user influence estimation apparatus of a social platform according to an embodiment of the present disclosure
  • FIG. 3b is another schematic structural diagram of a user influence estimation apparatus of a social platform according to an embodiment of the present invention.
  • the principles of the present invention operate using many other general purpose or special purpose computing, communication environments, or configurations.
  • Examples of well-known computing systems, environments, and configurations suitable for use with the present invention may include, but are not limited to, hand-held phones, personal computers, servers, multi-processor systems, microcomputer-based systems, mainframe computers, and A distributed computing environment, including any of the above systems or devices.
  • Embodiments of the present invention provide a method and apparatus for estimating user influence of a social platform.
  • FIG. 1 is a schematic diagram of a scenario for estimating a user influence of a social platform according to an embodiment of the present invention.
  • the scenario may include a user influence estimation device of the social platform, which is mainly referred to as an influence estimation device, and is mainly used for Obtaining user behavior data on a social platform, such as interactive information of a user on a social platform for a message posted by a friend, and/or interaction information of a user on the social platform for an advertisement served by the advertisement delivery system, and thereafter, according to the user Behavioral data, determining the influence transfer relationship between users, based on the influence transfer relationship, Estimate the ranking of influence of all users on the social platform. Finally, the influence of each user can be determined according to the influence ranking.
  • the scenario may further include a storage device, which is mainly used to store user behavior data on the social platform, such as interaction information of the user on the social platform for the message posted by the friend, and/or the user on the social platform for the advertisement delivery system.
  • the interactive information of the delivered advertisement, etc. is used for the influence estimation device to call the processing.
  • the scenario may also include a service device, such as an advertisement delivery device, for advertising the user's social platform based on the influence of the user's influence on the impact estimation device, and the like.
  • the description will be made from the perspective of the influence estimating device, and the influence estimating device may be specifically integrated in a network device such as a server or a gateway.
  • a method for estimating user influence of a social platform comprising: acquiring user behavior data on a social platform; determining, according to the user behavior data, an influence transfer relationship between the two users; and estimating the user to be social based on the influence transfer relationship
  • the ranking of influence on the platform determine the influence of each user based on the ranking of influence.
  • FIG. 1b is a schematic flowchart of a method for estimating a user influence of a social platform according to a first embodiment of the present invention.
  • the specific process may include:
  • step S101 user behavior data on the social platform is acquired.
  • step S102 based on the user behavior data, the influence transfer relationship between the two of the users is determined.
  • the social platform may specifically include a WeChat friend circle, a microblog, a QQ space, etc., and the user may share his or her mood, pay attention to the status of the friend, and learn some hot topics, news, and the like on the social platform.
  • a social platform may be configured to correspond to a database, and the influence estimating device may obtain user behavior data on the corresponding social platform from the databases;
  • the data of all the social platforms may be collated, and the influence estimation device may obtain user behavior data therefrom, which is not specifically limited herein.
  • determining the influence transfer relationship between the two users according to the user behavior data includes the following steps:
  • the influence transfer relationship between users can be described by the influence transfer matrix W ⁇ R ⁇ (n ⁇ n) , wherein the elements in the influence transfer matrix indicate the user between the two
  • the influence transfer relationship is the influence of one user on another.
  • step 1 there are many ways to generate an influence transfer matrix according to user behavior data (ie, step 1), which may specifically include:
  • the first interactive information is interactive information of a message posted by the user on the social platform for the personal release of the friend
  • the second interactive information is interaction information of the advertisement served by the user on the social platform on the advertisement delivery system
  • the user behavior data in the embodiment of the present invention may include interaction information (ie, first interaction information) of the user's personally posted message, and interactive information of the user's advertisement for the advertisement delivery system (ie, the second interaction).
  • the information estimating device generates an influence transfer matrix according to the first interaction information and the second interaction information, thereby determining an influence transfer relationship between the users.
  • the first interaction information may be specifically the number of times the user B comments (or likes) the message posted by the user A, and the second interaction information may be specifically that the user B continues after the user A comments (or likes) an advertisement.
  • the first interaction information may further include the number of interactions of the user B posting messages to all of the friends
  • the second interaction information may further include the user B in the The number of interactions of all friends after an engagement ad; in addition, it is also necessary to determine the importance weight value P of the friend information of the circle of friends, and the importance weight value Q of the friend interaction on the friend circle advertisement, thereby integrating the first interaction information.
  • the second interactive information importance weight value P and the importance weight value Q obtain the influence of the user A on the user B.
  • the influence transfer relationship between the other two users can also be determined by referring to the above manner, thereby constructing the influence transfer matrix; in addition, the specific values of the importance weight values P and Q in this embodiment may be It is determined according to the proportion of attention of the actual application scenario, and is not specifically limited herein.
  • step S103 based on the influence transfer relationship, the influence ranking of the user on the social platform is estimated.
  • step S104 the influence of each user is determined according to the influence ranking.
  • the present embodiment is to rank the influence of each user in the entire social network. Therefore, this embodiment can learn from The idea of the PageRank algorithm is to estimate the user's influence ranking.
  • PageRank is an algorithm designed by Larry Page and Sergey Brin to measure the importance of a particular page relative to other pages in the search engine. The result of the calculation is an important indicator of the ranking of the page in the google search results.
  • PageRank assumes that the user randomly selects a web page from all web pages to browse, and then continuously jumps through the hyperlinks on the web page. Once each page is reached, the user has two choices: end here or continue to select a link to browse.
  • the probability that the algorithm allows the user to continue browsing is d, and the user randomly selects one of the hyperlinks of the current page to continue browsing with equal probability. This can be considered as a random walk. After a number of such walks, the probability that each web page is accessed by the visiting user will converge to a stable value. This probability is the importance index of the web page and is used for page rank.
  • countless web pages in the Internet can constitute an oversized graph.
  • Each node in the graph is a web page, and the hyperlink is the edge in the graph.
  • PageRank uses a random walk to process the web page. Ranking.
  • Each node in the graph represents a user, and the interaction between users is seen as the edge in the graph.
  • PageRank's algorithm to the graph composed of social networks, ranking users and calculating user influence.
  • “estimating the ranking of the influence of the user on the social platform based on the influence transfer relationship” may include:
  • a random walk-based influence estimation algorithm can be designed on the social network. As time goes by, the user's influence ranking on the social platform will change accordingly.
  • the estimation algorithm before calculating the current influence ranking of the user, it is necessary to determine the ranking of the user's initial influence and the ranking of the influence of the user on the social platform at the previous moment (can be called historical influence ranking).
  • the “estimating the final impact ranking of the user according to the historical influence ranking and the current influence ranking” includes: if the historical influence ranking and the current influence difference satisfy the preset convergence condition , the current influence ranking is determined as the estimated result of the final influence ranking.
  • the influence ranking of the user on the social platform will converge to a fixed value over time, and this value is the estimated result of the final influence ranking, through the influence
  • the ranking results of the rankings can determine the impact value of each user on the social platform.
  • the user influence estimation method of the social platform first determines the influence transfer relationship between the two users according to the user behavior data on the social platform, and then estimates the user based on the influence transfer relationship.
  • the influence ranking on the social platform so that the influence of the user can be determined according to the influence ranking; since the user behavior data mainly reflects the interaction information of the user in the social activity, the program mainly determines the user between the user behavior data.
  • the influence transfer relationship is estimated based on the influence transfer relationship, so the user's social influence is greatly improved compared with the existing method of measuring the social influence of the user based on the friend coverage. Accuracy and credibility, which also improves the accuracy of information delivery on social platforms.
  • FIG. 2a is a schematic flowchart of a method for estimating a user influence of a social platform provided by the present invention.
  • the specific process may include:
  • step S201 the influence estimating device acquires user behavior data and according to user behavior Data, build an influence transfer matrix.
  • a network graph can be constructed.
  • Each node in the network graph represents a user, and the interaction between users is regarded as an edge in the network graph.
  • the influence transfer matrix W ⁇ R ⁇ (n ⁇ n) is constructed according to the first interaction information and the second interaction information, wherein the first interaction information is published by the user on the social platform for the friend personally.
  • the interactive information of the message which is the interactive information of the user on the social platform for the advertisement served by the advertisement delivery system.
  • the elements in the influence transfer matrix can be determined according to the following formula:
  • C ij is the number of comments (or likes) of user j's message to user i
  • a ij is the number of times user j continues to comment (or like) after user i comments (or likes) an ad.
  • k ⁇ N(u j ) is all neighbor friends of user j.
  • are the importance weights of the friend's user information and the friend interaction on the friend circle advertisement respectively. Since we are more concerned with the influence of users on advertising, there is usually ⁇ .
  • step S202 the influence estimating device generates an influence ranking estimation formula based on the preset webpage ranking algorithm and the influence transfer matrix.
  • step S203 the influence estimating device acquires the initialization influence ranking and the historical influence ranking of the user on the social platform.
  • step S204 the influence estimating device calculates the current influence ranking by using the influence ranking estimation formula based on the initial influence ranking and the historical influence ranking.
  • step S205 the influence estimating device determines the historical influence ranking and the Whether the difference in the pre-influence meets the preset convergence condition.
  • step S206 If yes, go to step S206, if no, go back to step S204;
  • step S206 the influence estimating device determines the current influence ranking as the estimation result of the user influence ranking and outputs.
  • the step S202 to the step S206 may be specifically:
  • the influence transfer matrix describes the influence of the user i on the user j, that is, the probability that the user j will focus on the message of the user i. That is, w(i, j) describes the influence transfer relationship between the two users.
  • w(i, j) describes the influence transfer relationship between the two users.
  • PageRank a random motion estimation algorithm based on random walk (the influence ranking estimation formula) is designed on the social network G.
  • the calculation formula of the algorithm is as follows:
  • FIG. 2b For the interaction of friends in this embodiment, assuming WeChat friends circle
  • the interactive network consists of 4 users, and the interaction between them is shown in Figure 2b.
  • nodes u1, u2, u3, and u4 represent 4 users
  • directed edges represent interactions between users.
  • the directed edge u4->u1 indicates the behavior of the user u4 to the user u1.
  • the two numbers on the side indicate that the user u4 has liked the message posted by the user u1 twice, and the advertisement that comments on the user u1 has one time. Enter the comment.
  • the influence transfer matrix is constructed by considering the interaction record of the user in the circle of friends advertisement and the personal information of the circle of friends, and the random design matrix is designed.
  • the walk algorithm can realize the user influence estimation; further, the user influence estimation result is applied to the advertisement placement in the circle of friends, and the advertisement can be preferentially placed to the high-impact users, and the advertisement system receives the comments of the users or After praising and then serving to friends of high-impact users, you can greatly improve the interactive rate of advertising and achieve better advertising efficiency.
  • the user influence estimation method of the social platform first determines the influence transfer relationship between the two users according to the user behavior data on the social platform, and then estimates the user based on the influence transfer relationship.
  • Ranking influence on social platforms thus The influence of the user is determined according to the influence ranking; since the user behavior data mainly reflects the interaction information of the user in the social activity, the scheme mainly determines the influence transfer relationship between the users according to the user behavior data, and based on the influence transfer
  • the relationship estimates the influence of the user, so the accuracy and credibility of the user's social influence estimation is greatly improved compared to the existing method of measuring the social influence of the user based on the degree of friend coverage. Improve the accuracy of information delivery on social platforms.
  • the embodiment of the present invention further provides an apparatus for estimating the user influence force based on the social platform.
  • the meaning of the noun is the same as the method for estimating the user influence of the social platform mentioned above. For specific implementation details, refer to the description in the method embodiment.
  • FIG. 3a is a schematic structural diagram of a user influence estimation apparatus of a social platform according to an embodiment of the present invention.
  • the apparatus may include an obtaining unit 301, a first determining unit 302, an estimating unit 303, and a second determining unit 304.
  • the obtaining unit 301 is configured to acquire user behavior data on the social platform.
  • the first determining unit 302 is configured to determine, according to the user behavior data, an influence transfer relationship between the two.
  • the social platform may specifically include a WeChat friend circle, a microblog, a QQ space, etc., and the user may share his or her mood, pay attention to the status of the friend, and learn some hot topics, news, and the like on the social platform.
  • a social platform may be configured to correspond to a database, and the impact estimating device may obtain user behavior data on the corresponding social platform from the databases; in some embodiments, all social platforms may be The data is sorted, and the influence estimating device can obtain user behavior data therefrom, which is not specifically limited herein.
  • this embodiment is to rank the influence of each user in the entire social network. Therefore, this embodiment can learn PageRank (page rank).
  • PageRank page rank
  • the idea of the algorithm is to estimate the user influence ranking.
  • Each node in the graph is a web page.
  • the hyperlink is the edge in the graph.
  • PageRank ranks the web page through a random walk. Based on this, in the social network, we can also form an oversized graph.
  • Each node in the graph represents a user, and the interaction between users is seen as the edge in the graph.
  • PageRank's algorithm to the graph composed of social networks, ranking users and calculating user influence.
  • the estimating unit 303 is configured to estimate the influence ranking of the user on the social platform based on the influence transfer relationship; the second determining unit 304 is configured to determine the influence of each user according to the influence ranking.
  • FIG. 3b is a schematic structural diagram of a user influence estimation apparatus of a social platform according to an embodiment of the present invention; wherein the first determining unit 302 may specifically include:
  • a matrix generation subunit 3021 configured to generate an influence force transfer matrix according to the user behavior data
  • the first determining subunit 3022 is configured to determine an influence transfer relationship between the two users according to the influence transfer matrix.
  • the influence transfer relationship between users can be described by the influence transfer matrix W ⁇ R ⁇ (n ⁇ n) , wherein the elements in the influence transfer matrix indicate the user between the two
  • the influence transfer relationship is the influence of one user on another.
  • matrix generation subunit 3021 may be specifically configured to:
  • the interaction information is interaction information of the advertisements of the user on the advertisement platform for the advertisement delivery system; and the influence transfer matrix is generated according to the first interaction information and the second interaction information.
  • the user behavior data in the embodiment of the present invention may include interaction information (ie, first interaction information) of the user's personally posted message, and interactive information of the user's advertisement for the advertisement delivery system (ie, the second interaction).
  • the information estimating device generates an influence transfer matrix according to the first interaction information and the second interaction information, thereby determining an influence transfer relationship between the users.
  • the first interaction information may be specifically the number of times the user B comments (or likes) the message posted by the user A for the WeChat friend circle, and the second interaction The information may be specifically the number of times User B continues to comment (or like) after User A comments (or likes) an advertisement.
  • the first interaction information may further include the number of interactions of the user B posting messages to all of the friends
  • the second interaction information may further include the user B in the The number of interactions of all friends after an engagement ad; in addition, it is also necessary to determine the importance weight value P of the friend information of the circle of friends, and the importance weight value Q of the friend interaction on the friend circle advertisement, thereby integrating the first interaction information.
  • the second interactive information importance weight value P and the importance weight value Q obtain the influence of the user A on the user B.
  • the influence transfer relationship between the other two users can also be determined by referring to the above manner, thereby constructing the influence transfer matrix; in addition, the specific values of the importance weight values P and Q in this embodiment may be It is determined according to the proportion of attention of the actual application scenario, and is not specifically limited herein.
  • the estimating unit 303 may specifically include:
  • the obtaining sub-unit 3031 is configured to obtain an initializing influence ranking and a historical influence ranking of the user on the social platform, where the historical influence ranking is a ranking of the influence of the user on the social platform in the previous moment;
  • the estimating sub-unit 3032 is configured to estimate a current influence ranking by using a preset webpage ranking algorithm, based on the influence transfer relationship, the initial influence ranking, and the historical influence ranking, the current influence ranking The ranking of the influence of the user on the social platform at the current moment.
  • a random walk-based influence estimation algorithm can be designed on the social network. As time goes by, the user's influence ranking on the social platform will change accordingly.
  • the estimation algorithm before calculating the current influence ranking of the user, it is necessary to determine the ranking of the user's initial influence and the ranking of the influence of the user on the social platform at the previous moment (can be called historical influence ranking).
  • the estimating sub-unit 3032 further needs to analyze the current influence ranking to determine the final influence ranking of the user, for example, may also be used to rank according to the historical influence ranking and the current influence ranking. Estimating the final impact ranking of the user, determining the final impact ranking as the ranking of influence of the user on the social platform.
  • the estimating subunit 3032 may be further configured to: if the historical influence ranking and the current influence difference meet the preset convergence condition, determine the current influence ranking as the final impact ranking estimation. result.
  • the influence ranking of the user on the social platform will converge to a fixed value over time, and this value is the estimated result of the final influence ranking, through the influence
  • the ranking results of the rankings can determine the impact value of each user on the social platform.
  • the foregoing units may be implemented as a separate entity, or may be implemented in any combination, and may be implemented as the same or a plurality of entities.
  • the foregoing method embodiments and details are not described herein.
  • the user influence estimating device of the social platform may be specifically integrated in a network device such as a server or a gateway.
  • the user influence estimation device of the social platform first determines the influence transfer relationship between the two users according to the user behavior data on the social platform, and then estimates the user based on the influence transfer relationship.
  • the influence ranking on the social platform so that the influence of the user can be determined according to the influence ranking; since the user behavior data mainly reflects the interaction information of the user in the social activity, the program mainly determines the user between the user behavior data.
  • the influence transfer relationship is estimated based on the influence transfer relationship, so the user's social influence is greatly improved compared with the existing method of measuring the social influence of the user based on the friend coverage. Accuracy and credibility, which also improves the accuracy of information delivery on social platforms.
  • the device influence estimating device of the social platform provided by the embodiment of the present invention is, for example, a computer, a tablet computer, a mobile phone with touch function, or the like, the user influence estimating device of the social platform and the social in the above embodiment.
  • the user impact estimation method of the platform belongs to the same concept, and any method provided in the embodiment of the user influence estimation method of the social platform may be run on the user influence estimation device of the social platform, and the specific implementation process thereof is detailed.
  • the user influence estimation method embodiment of the social platform is not described here.
  • a common tester in the field can understand all or part of the process for implementing the user influence estimation method of the social platform according to the embodiment of the present invention.
  • This is accomplished by a computer program controlling the associated hardware, which may be stored in a computer readable storage medium, such as in a memory of the terminal, and executed by at least one processor within the terminal, during execution
  • a flow may be included in an embodiment of a user impact estimation method as described for the social platform.
  • the storage medium may be a magnetic disk, an optical disk, a read only memory (ROM, Read Only Memory), Random access memory (RAM, Random Access Memory).
  • each functional module may be integrated into one processing chip, or each module may exist separately physically, or two or more modules may be integrated in the module.
  • a module The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.
  • the integrated module if implemented in the form of a software functional module and sold or used as a standalone product, may also be stored in a computer readable storage medium, such as a read only memory, a magnetic disk or an optical disk, etc. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Computing Systems (AREA)
  • Finance (AREA)
  • Development Economics (AREA)
  • Accounting & Taxation (AREA)
  • Tourism & Hospitality (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Primary Health Care (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

一种社交平台的用户影响力估算方法、装置及计算机存储介质,其中该方法包括:获取社交平台上的用户行为数据(S101);根据用户行为数据,确定用户两两之间的影响力转移关系(S102);基于影响力转移关系,估算用户在社交平台上的影响力排名(S103);根据影响力排名确定各用户的影响力(S104)。由于用户行为数据主要体现了用户在社交活动中的互动信息,而该方法主要根据用户行为数据确定用户之间的影响力转移关系,并基于影响力转移关系对用户的影响力进行估算,因此相对于现有单靠基于好友覆盖程度来度量用户的社交影响力的方式,大大的提高了用户社交影响力估算的准确度和可信度,从而也提高了社交平台上信息投放的精确度。

Description

社交平台的用户影响力估算方法、装置及计算机存储介质
本专利申请要求2016年01月07日提交的中国专利申请号为201610009657.3,申请人为腾讯科技(深圳)有限公司,发明名称为“一种社交平台的用户影响力估算方法及装置”的优先权,该申请的全文以引用的方式并入本申请中。
技术领域
本发明涉及通信技术领域,尤其涉及一种社交平台的用户影响力估算方法及装置。
背景技术
随着互联网技术的发展,各种社交应用也越来越广泛,在社交网络的平台上,人们可以分享自己的心情、关注朋友的状态以及了解一些热门话题、新闻等。社交应用中所涉及的大量用户数据,比如用户的喜好、社交活动和用户社交影响力(可简称用户影响力)等等,对于信息的投放有着极大的意义。
目前传统技术上,用户影响力的确定,一般基于人脉网络。在社交网络中用户可以将自己喜爱的人加为好友,甚至加为密友,因此,基于人脉网络的影响力计算方式,就是利用用户的好友覆盖程度来计算,一个用户的拥有的好友数目越多,其社交影响力就越高。用户影响力描述了一个用户影响其他用户的能力,在社交网络领域(如微信朋友圈等),用户影响力可以用该用户收到的关注度来度量,受关注度越高,其社交影响力就越大。
现有的用户影响力估算方案虽然在一定程度中可以估算出用户的社交影响力,但是如果用户的好友数目虽然很多,但经常联系的却很少,则单单基于好友覆盖程度来度量用户的社交影响力,会使得计算出的用户社交 影响力准确度和可信度不高,从而导致社交平台上的信息投放也不够精确。
发明内容
本发明的目的在于提供一种社交平台的用户影响力估算方法及装置,旨在提高用户社交影响力计算的准确度和可信度,从而提高社交平台上信息投放的精确度。
为解决上述技术问题,本发明实施例提供以下技术方案:
一种社交平台的用户影响力估算方法,包括:
获取社交平台上的用户行为数据;
根据所述用户行为数据,确定用户两两之间的影响力转移关系;
基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;
根据所述影响力排名确定各用户的影响力。
为解决上述技术问题,本发明实施例还提供以下技术方案:
一种社交平台的用户影响力估算装置,包括:
获取单元,用于获取社交平台上的用户行为数据;
第一确定单元,用于根据所述用户行为数据,确定用户两两之间的影响力转移关系;
估算单元,用于基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;
第二确定单元,用于根据所述影响力排名确定各用户的影响力。
一种计算机存储介质,该存储介质包括一组指令,当执行所述指令时,引起至少一个处理器执行包括以下的操作:
获取社交平台上的用户行为数据;
根据所述用户行为数据,确定用户两两之间的影响力转移关系;
基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;
根据所述影响力排名确定各用户的影响力。
相对于现有技术,本发明实施例,首先根据社交平台上的用户行为数据,确定用户两两之间的影响力转移关系,其后基于影响力转移关系,估算用户在社交平台上的影响力排名,从而可以根据影响力排名确定用户的影响力;由于用户行为数据主要体现了用户在社交活动中的互动信息,而该方案主要根据用户行为数据确定用户之间的影响力转移关系,并基于影响力转移关系对用户的影响力进行估算的,因此相对于现有单靠基于好友覆盖程度来度量用户的社交影响力的方式,大大的提高了用户社交影响力估算的准确度和可信度,从而也提高了社交平台上信息投放的精确度。
附图说明
下面结合附图,通过对本发明的具体实施方式详细描述,将使本发明的技术方案及其它有益效果显而易见。
图1a是本发明第实施例提供的社交平台的用户影响力估算方法的场景示意图;
图1b是本发明实施例提供的社交平台的用户影响力估算方法的流程示意图;
图2a为本发明实施例提供的社交平台的用户影响力估算方法的流程示意图;
图2b为本发明实施例提供的社交平台的用户影响力估算方法的应用示意图;
图3a为本发明实施例提供的社交平台的用户影响力估算装置的结构示意图;
图3b为本发明实施例提供的社交平台的用户影响力估算装置的另一结构示意图。
具体实施方式
请参照图式,其中相同的组件符号代表相同的组件,本发明的原理是以实施在一适当的运算环境中来举例说明。以下的说明是基于所例示的本发明具体实施例,其不应被视为限制本发明未在此详述的其它具体实施例。
在以下的说明中,本发明的具体实施例将参考由一部或多部计算机所执行的步骤及符号来说明,除非另有述明。因此,这些步骤及操作将有数次提到由计算机执行,本文所指的计算机执行包括了由代表了以一结构化形式中的数据的电子信号的计算机处理单元的操作。此操作转换该数据或将其维持在该计算机的内存系统中的位置处,其可重新配置或另外以本领域测试人员所熟知的方式来改变该计算机的运作。该数据所维持的数据结构为该内存的实体位置,其具有由该数据格式所定义的特定特性。但是,本发明原理以上述文字来说明,其并不代表为一种限制,本领域测试人员将可了解到以下所述的多种步骤及操作亦可实施在硬件当中。
本发明的原理使用许多其它泛用性或特定目的运算、通信环境或组态来进行操作。所熟知的适合用于本发明的运算系统、环境与组态的范例可包括(但不限于)手持电话、个人计算机、服务器、多处理器系统、微电脑为主的系统、主架构型计算机、及分布式运算环境,其中包括了任何的上述系统或装置。
本发明实施例提供一种社交平台的用户影响力估算方法及装置。
参见图1a,该图为本发明实施例所提供的社交平台的用户影响力估算方法的场景示意图,该场景中,可以包括社交平台的用户影响力估算装置,简称影响力估算装置,主要用于获取社交平台上的用户行为数据,比如,社交平台上用户对于好友个人发布的消息的互动信息,和/或社交平台上用户对于广告投放系统投放的广告的互动信息等,其后,根据这些用户行为数据,确定用户两两之间的影响力转移关系,并基于该影响力转移关系, 估算所有用户在社交平台上的影响力排名,最后,可以根据所述影响力排名确定各用户的影响力。
此外,该场景中,还可以包括存储设备,主要用于存储社交平台上的用户行为数据,如社交平台上用户对于好友个人发布的消息的互动信息,和/或社交平台上用户对于广告投放系统投放的广告的互动信息等,供影响力估算装置调用处理。当然,该场景中还可以包括业务设备,如广告投放设备,用于根据影响力估算装置输出的用户影响力,向用户社交平台投放广告,等等。
以下将分别进行详细说明。
在本实施例中,将从影响力估算装置的角度进行描述,该影响力估算装置具体可以集成在服务器或网关等网络设备中。
一种社交平台的用户影响力估算方法,包括:获取社交平台上的用户行为数据;根据该用户行为数据,确定用户两两之间的影响力转移关系;基于影响力转移关系,估算用户在社交平台上的影响力排名;根据影响力排名确定各用户的影响力。
请参阅图1b,图1b是本发明第一实施例提供的社交平台的用户影响力估算方法的流程示意图,具体流程可以包括:
在步骤S101中,获取社交平台上的用户行为数据。
在步骤S102中,根据所述用户行为数据,确定用户两两之间的影响力转移关系。
本发明实施例中,社交平台可以具体包括微信朋友圈、微博、QQ空间等,用户可以在社交平台上分享自己的心情、关注朋友的状态以及了解一些热门话题、新闻等。
在某些实施方式中,可以设置一个社交平台对应一个数据库,影响力估算装置可以从这些数据库中,获取对应的社交平台上的用户行为数据; 在某些实施方式中,可以将所有社交平台的数据进行整理,影响力估算装置可以从中获取用户行为数据,此处不作具体限定。
可具体的,比如“根据所述用户行为数据,确定用户两两之间的影响力转移关系”包括如下步骤:
1、根据获取到的用户行为数据,生成影响力转移矩阵;
2、根据影响力转移矩阵,确定用户两两之间的影响力转移关系。
也就是说,比如,可以将用户之间的影响力转移关系用影响力转移矩阵W∈R^(n×n)进行描述,其中,该影响力转移矩阵中的元素指示用户两两之间的影响力转移关系,即指示一个用户对另一个用户的影响力。
可以理解的是,在社交平台(如微信)中,用户影响力是用户改变和吸引其他用户行为的能力。影响力高的用户受到其好友的关注更高,发表的信息、获取的评论和点赞个数更多,信息观点传播的速度更快。
进一步的,“根据用户行为数据,生成影响力转移矩阵”(即步骤1)的方式有很多,其中可具体包括:
11、基于用户行为数据确定第一互动信息和第二互动信息;
其中,该第一互动信息为社交平台上用户对于好友个人发布的消息的互动信息,该第二互动信息为社交平台上用户对于广告投放系统投放的广告的互动信息;
12、根据第一互动信息和第二互动信息生成影响力转移矩阵。
也就是说,本发明实施例中的用户行为数据可以包括用户对于好友个人发布的消息的互动信息(即第一互动信息),以及用户对于广告投放系统投放的广告的互动信息(即第二互动信息),影响力估算装置根据第一互动信息和第二互动信息来生成影响力转移矩阵,从而确定用户之间的影响转移关系。
比如,若需要确定用户A对用户B的影响力,则针对于微信朋友圈, 第一互动信息可以具体为用户B对用户A所发布的消息的评论(或者点赞)的次数,第二互动信息可以具体为用户B在用户A评论(或者点赞)了某个广告后继续评论(或者点赞)的次数。
进一步的,生成影响力转移矩阵过程中,还需要确定如下参数,比如:第一互动信息还可以包括用户B对其所有好友个人发布消息的互动次数,第二互动信息还可以包括用户B在其所有好友个人于某个互动广告后的互动次数;另外,还需要确定设定朋友圈用户信息重要性权重值P,以及朋友圈广告上好友互动的重要性权重值Q,从而综合第一互动信息、第二互动信息重要性权重值P、重要性权重值Q,得到用户A对用户B的影响力。
容易想到的是,其他两两用户之间的影响力转移关系也可以参照上述方式进行确定,从而构造出影响力转移矩阵;另外,本实施例的重要性权重值P和Q的具体取值可以根据实际应用场景的关注比例进行确定,此处不作具体限定。
在步骤S103中,基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名。
在步骤S104中,根据所述影响力排名确定各用户的影响力。
可以理解的是,由于影响力转移矩阵是描述用户两两之间的影响力转移关系,而本实施例是要对每个用户在整个社交网络中的影响力排名,因此,本实施例可以借鉴PageRank(网页排名)算法的思路,对用户影响力排名进行估算。
以下对PageRank算法作简单说明:
PageRank是Larry Page和Sergey Brin设计的用来衡量特定网页相对于搜索引擎中其他网页的重要性的算法,其计算结果作为google搜索结果中网页排名的重要指标。
由于网页之间通过超链接相互连接,互联网上不计其数的网页就构成 了一张超大的图。PageRank假设用户从所有网页中随机选择一个网页进行浏览,然后通过超链接在网页直接不断跳转。到达每个网页后,用户有两种选择:到此结束或者继续选择一个链接浏览。该算法令用户继续浏览的概率为d,用户以相等的概率在当前页面的所有超链接中随机选择一个继续浏览。这可以认为是一个随机游走的过程。当经过多次这样的游走之后,每个网页被访问用户访问到的概率就会收敛到一个稳定值。这个概率就是网页的重要性指标,被用于网页排名。
如上所描述,互联网中不计其数的网页可以构成一个超大的图,图中的每一个节点是一个网页,超链接是图中的边,在这个图上PageRank通过随机游走的过程来对网页进行排名。基于此,在社交网络中,我们也可以构成一个超大的图,图中的每一个节点代表一个用户,用户之间的互动关系看成图中的边。同样的,我们可以把PageRank的算法应用到社交网络构成的图上,对用户进行排名、计算用户的影响力。
在本实施例中,“基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名”可以包括:
a、获取用户在所述社交平台上的初始化影响力排名和历史影响力排名,所述历史影响力排名为上一时刻用户在所述社交平台上的影响力排名;
b、通过预设网页排名算法,基于所述影响力转移关系、所述初始化影响力排名以及历史影响力排名,估算当前影响力排名,所述当前影响力排名为当前时刻用户在所述社交平台上的影响力排名。
可以理解的是,基于PageRank思路,在社交网络上可以设计一个基于随机游走的影响力预估算法,随着时间的推移,用户在社交平台上的影响力排名会随之变化,该影响力预估算法中,在计算用户当前影响力排名之前,需要确定用户的初始化影响力排名以及上一时刻用户在所述社交平台上的影响力排名(可称历史影响力排名)。
更进一步的,在“估算当前影响力排名”之后,还需要对当前影响力排名进行分析,以确定用户的最终影响力排名,比如:
c、根据所述历史影响力排名以及所述当前影响力排名,估算用户最终影响力排名;
d、将该最终影响力排名确定为用户在所述社交平台上的影响力排名。
可具体的,“根据所述历史影响力排名以及所述当前影响力排名,估算用户最终影响力排名”,包括:若所述历史影响力排名以及所述当前影响力的差别满足预设收敛条件,则将当前影响力排名确定为最终影响力排名的估算结果。
也就是说,对于所有用户来说,随着时间的推移,用户在所述社交平台上的影响力排名会收敛到一个固定的值,而这个值就是最终影响力排名的估算结果,通过该影响力排名估算结果,可以确定出各用户在该社交平台上的影响力值。
由上述可知,本实施例提供的社交平台的用户影响力估算方法,首先根据社交平台上的用户行为数据,确定用户两两之间的影响力转移关系,其后基于影响力转移关系,估算用户在社交平台上的影响力排名,从而可以根据影响力排名确定用户的影响力;由于用户行为数据主要体现了用户在社交活动中的互动信息,而该方案主要根据用户行为数据确定用户之间的影响力转移关系,并基于影响力转移关系对用户的影响力进行估算的,因此相对于现有单靠基于好友覆盖程度来度量用户的社交影响力的方式,大大的提高了用户社交影响力估算的准确度和可信度,从而也提高了社交平台上信息投放的精确度。
参阅图2a,图2a为本发明提供的社交平台的用户影响力估算方法的流程示意图,具体流程可以包括:
在步骤S201中,影响力估算装置获取用户行为数据,并根据用户行为 数据,构建影响力转移矩阵。
首先,基于PageRank思想,在社交网络中,可以构成一个网络图,网络图中的每一个节点代表一个用户,用户之间的互动关系看成网络图中的边。
比如,在微信朋友圈的社交平台上,将用户之间的互动构成一个庞大的网络G={V,E},其中节点为V={u1,u2,…,un},n为用户个数,边为E={eij|ui和uj为好友}。在这个网络结构的基础上,根据第一互动信息和第二互动信息,构建影响力转移矩阵W∈R^(n×n),其中,该第一互动信息为社交平台上用户对于好友个人发布的消息的互动信息,该第二互动信息为社交平台上用户对于广告投放系统投放的广告的互动信息。
更进一步的,该影响力转移矩阵中的元素可根据如下公式进行确定:
Figure PCTCN2017070503-appb-000001
其中Cij为用户j对用户i所发布的消息的评论(或者点赞)次数,Aij为用户j在用户i评论(或者点赞)了某个广告后继续评论(或者点赞)的次数,k∈N(uj)为用户j的所有邻居好友。α,β分别为朋友圈用户信息和朋友圈广告上好友互动的重要性权重值。由于我们更关注用户在广告上的影响力,因此一般会有α<β。
在步骤S202中,影响力估算装置基于预设网页排名算法及该影响力转移矩阵,生成影响力排名估算公式。
在步骤S203中,影响力估算装置获取用户在所述社交平台上的初始化影响力排名和历史影响力排名。
在步骤S204中,影响力估算装置基于所述初始化影响力排名和历史影响力排名,采用影响力排名估算公式计算当前影响力排名。
在步骤S205中,影响力估算装置判断所述历史影响力排名以及所述当 前影响力的差别是否满足预设收敛条件。
若是,则执行步骤S206,若否,则返回执行步骤S204;
在步骤S206中,影响力估算装置将当前影响力排名确定为用户影响力排名的估算结果并输出。
其中,所述步骤S202至步骤S206可具体为:
由于,影响力转移矩阵中的元素w(i,j)描述的是用户i对用户j的影响力,也就是用户j会将注意力放在用户i的消息上的概率。即,w(i,j)描述的是用户两两之间的影响力转移关系,本发明实施例中,我们需要得到每个用户在整个社交网络中的影响力排名。因此,借鉴PageRank的思路,在社交网络G上设计一个基于随机游走的影响力预估算法(即影响力排名估算公式),其中算法的计算公式如下所示:
I(t+1)=bWIt+(1-b)I0    (2)
其中It∈R^(1×n)是一个向量,描述t时刻所有用户的影响力排名;当t=0时,I0的每一个元素值等于1/n;b为可调节超参数,根据经验值进行设定,一般设置在0.8-0.9之间。
基于该公式(2)可知,若需要知道当前影响力排名(即I(t+1)),需要先获取用户在所述社交平台上的初始化影响力排名(即I0)和历史影响力排名(即上一时刻的影响力排名It),随后,判断所述历史影响力排名以及所述当前影响力的差别是否满足预设收敛条件,若是,则将当前影响力排名确定为用户影响力排名的估算结果并输出。
也就是说,在公式(2)中,针对每一随机用户,从自己节点开始带着1/n的影响力值沿着网络中的边,以矩阵W中的影响力转移概率访问其邻居节点并将影响力按比例传递给邻居。当随着时间的推移,每个用户的影响力值It会收敛到一个固定的值,这个值就是最终的用户影响力排名。
为了更好的理解本发明技术方案,以下以一具体应用例进行分析说明:
可一并参考图2b,为该实施例中的好友互动的示意,假设微信朋友圈 的互动网络由4个用户构成,他们之间的互动情况如图2b所示,其中,节点u1,u2,u3,u4表示4个用户,有向边表示用户之间的互动。
例如,有向边u4->u1表示用户u4对用户u1的行为,边上的两个数字分别表示用户u4对用户u1发布的消息有2次点赞,对用户u1评论的广告有1次跟进的评论。
其中在公式(1)中,为了便于计算,本实施例中可以设置α=0.5,β=0.5,其后,基于公式(1)计算出影响力转移矩阵为:
Figure PCTCN2017070503-appb-000002
随后,基于公式(2),先初始化I0=(0.25,0.25,0.25,0.25),并可以设置b=0.85。接着,将W,I0,b代入进行迭代运算,从而当随着时间的推移,用户的影响力排名It会收敛到一个固定的值,这个值就是最终的用户影响力排名,由迭代计算可知,获得用户的最后的影响力排名It=(1.29,1.33,0.87,1.13),从而可以看出用户u2的影响力最大,u3的影响力最小。
若将本发明实施提供的社交平台的用户影响力估算方法,应用于微信用户影响力计算,则结合考虑用户在朋友圈广告和朋友圈个人信息上的互动记录构建影响力转移矩阵,并设计随机游走算法可以实现用户影响力预估;进一步的,将用户影响力预估结果应用于朋友圈中的广告投放,可以优先投放广告给高影响力用户,待广告系统收到这些用户的评论或者点赞后再投放给高影响力用户的好友,可以大大提高广告的互动率,达到更好的广告效益。
由上述可知,本实施例提供的社交平台的用户影响力估算方法,首先根据社交平台上的用户行为数据,确定用户两两之间的影响力转移关系,其后基于影响力转移关系,估算用户在社交平台上的影响力排名,从而可 以根据影响力排名确定用户的影响力;由于用户行为数据主要体现了用户在社交活动中的互动信息,而该方案主要根据用户行为数据确定用户之间的影响力转移关系,并基于影响力转移关系对用户的影响力进行估算的,因此相对于现有单靠基于好友覆盖程度来度量用户的社交影响力的方式,大大的提高了用户社交影响力估算的准确度和可信度,从而也提高了社交平台上信息投放的精确度。
为便于更好的实施本发明实施例提供的社交平台的用户影响力估算方法,本发明实施例还提供一种基于上述社交平台的用户影响力估算方法的装置。其中名词的含义与上述社交平台的用户影响力估算的方法中相同,具体实现细节可以参考方法实施例中的说明。
请参阅图3a,图3a为本发明实施例提供的社交平台的用户影响力估算装置的结构示意图,该装置可以包括获取单元301、第一确定单元302、估算单元303以及第二确定单元304。
其中,所述获取单元301,用于获取社交平台上的用户行为数据;第一确定单元302,用于根据所述用户行为数据,确定用户两两之间的影响力转移关系。
本发明实施例中,社交平台可以具体包括微信朋友圈、微博、QQ空间等,用户可以在社交平台上分享自己的心情、关注朋友的状态以及了解一些热门话题、新闻等。
在某些实施方式中,可以设置一个社交平台对应一个数据库,影响力估算装置可以从这些数据库中,获取对应的社交平台上的用户行为数据;在某些实施方式中,可以将所有社交平台的数据进行整理,影响力估算装置可以从中获取用户行为数据,此处不作具体限定。
可以理解的是,在社交平台(如微信)中,用户影响力是用户改变和吸引其他用户行为的能力。影响力高的用户受到其好友的关注更高,发表 的信息获取的评论和点赞个数更多,信息观点传播的速度更快。
由于影响力转移矩阵是描述用户两两之间的影响力转移关系,而本实施例是要对每个用户在整个社交网络中的影响力排名,因此,本实施例可以借鉴PageRank(网页排名)算法的思路,对用户影响力排名进行估算。
互联网中不计其数的网页可以构成一个超大的图,图中的每一个节点是一个网页,超链接是图中的边,在这个图上PageRank通过随机游走的过程来对网页进行排名。基于此,在社交网络中,我们也可以构成一个超大的图,图中的每一个节点代表一个用户,用户之间的互动关系看成图中的边。同样的,我们可以把PageRank的算法应用到社交网络构成的图上,对用户进行排名、计算用户的影响力。
估算单元303,用于基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;第二确定单元304,用于根据所述影响力排名确定各用户的影响力。
请一并参阅图3b,图3b为本发明实施例提供的社交平台的用户影响力估算装置的结构示意图;其中所述第一确定单元302,可以具体包括:
1、矩阵生成子单元3021,用于根据所述用户行为数据,生成影响力转移矩阵;
2、第一确定子单元3022,用于根据所述影响力转移矩阵,确定用户两两之间的影响力转移关系。
也就是说,比如,可以将用户之间的影响力转移关系用影响力转移矩阵W∈R^(n×n)进行描述,其中,该影响力转移矩阵中的元素指示用户两两之间的影响力转移关系,即指示一个用户对另一个用户的影响力。
进一步的,所述矩阵生成子单元3021,可以具体用于:
基于所述用户行为数据确定第一互动信息和第二互动信息,所述第一互动信息为社交平台上用户对于好友个人发布的消息的互动信息,所述第 二互动信息为社交平台上用户对于广告投放系统投放的广告的互动信息;根据所述第一互动信息和所述第二互动信息生成影响力转移矩阵。
也就是说,本发明实施例中的用户行为数据可以包括用户对于好友个人发布的消息的互动信息(即第一互动信息),以及用户对于广告投放系统投放的广告的互动信息(即第二互动信息),影响力估算装置根据第一互动信息和第二互动信息来生成影响力转移矩阵,从而确定用户之间的影响转移关系。
比如,若需要确定用户A对用户B的影响力,则针对于微信朋友圈,第一互动信息可以具体为用户B对用户A所发布的消息的评论(或者点赞)的次数,第二互动信息可以具体为用户B在用户A评论(或者点赞)了某个广告后继续评论(或者点赞)的次数。
进一步的,生成影响力转移矩阵过程中,还需要确定如下参数,比如:第一互动信息还可以包括用户B对其所有好友个人发布消息的互动次数,第二互动信息还可以包括用户B在其所有好友个人于某个互动广告后的互动次数;另外,还需要确定设定朋友圈用户信息重要性权重值P,以及朋友圈广告上好友互动的重要性权重值Q,从而综合第一互动信息、第二互动信息重要性权重值P、重要性权重值Q,得到用户A对用户B的影响力。
容易想到的是,其他两两用户之间的影响力转移关系也可以参照上述方式进行确定,从而构造出影响力转移矩阵;另外,本实施例的重要性权重值P和Q的具体取值可以根据实际应用场景的关注比例进行确定,此处不作具体限定。
基于此,在本实施例中,所述估算单元303可以具体包括:
a、获取子单元3031,用于获取用户在所述社交平台上的初始化影响力排名和历史影响力排名,所述历史影响力排名为上一时刻用户在所述社交平台上的影响力排名;
b、估算子单元3032,用于通过预设网页排名算法,基于所述影响力转移关系、所述初始化影响力排名以及所述历史影响力排名,估算当前影响力排名,所述当前影响力排名为当前时刻用户在所述社交平台上的影响力排名。
可以理解的是,基于PageRank思路,在社交网络上可以设计一个基于随机游走的影响力预估算法,随着时间的推移,用户在社交平台上的影响力排名会随之变化,该影响力预估算法中,在计算用户当前影响力排名之前,需要确定用户的初始化影响力排名以及上一时刻用户在所述社交平台上的影响力排名(可称历史影响力排名)。
更进一步的,所述估算子单元3032,还需要对当前影响力排名进行分析,以确定用户的最终影响力排名,比如,还可以用于根据所述历史影响力排名以及所述当前影响力排名,估算用户最终影响力排名,将所述最终影响力排名确定为用户在所述社交平台上的影响力排名。
可具体的,所述估算子单元3032,还可以用于若所述历史影响力排名以及所述当前影响力的差别满足预设收敛条件,则将当前影响力排名确定为最终影响力排名的估算结果。
也就是说,对于所有用户来说,随着时间的推移,用户在所述社交平台上的影响力排名会收敛到一个固定的值,而这个值就是最终影响力排名的估算结果,通过该影响力排名估算结果,可以确定出各用户在该社交平台上的影响力值。
具体实施时,以上各个单元可以作为独立的实体来实现,也可以进行任意组合,作为同一或若干个实体来实现,以上各个单元的具体实施可参见前面的方法实施例,在此不再赘述。
该社交平台的用户影响力估算装置具体可以集成在服务器或网关等网络设备中。
由上述可知,本实施例提供的社交平台的用户影响力估算装置,首先根据社交平台上的用户行为数据,确定用户两两之间的影响力转移关系,其后基于影响力转移关系,估算用户在社交平台上的影响力排名,从而可以根据影响力排名确定用户的影响力;由于用户行为数据主要体现了用户在社交活动中的互动信息,而该方案主要根据用户行为数据确定用户之间的影响力转移关系,并基于影响力转移关系对用户的影响力进行估算的,因此相对于现有单靠基于好友覆盖程度来度量用户的社交影响力的方式,大大的提高了用户社交影响力估算的准确度和可信度,从而也提高了社交平台上信息投放的精确度。
在上述实施例中,对各个实施例的描述都各有侧重,某个实施例中没有详述的部分,可以参见上文针对社交平台的用户影响力估算方法的详细描述,此处不再赘述。
本发明实施例提供的所述社交平台的用户影响力估算装置,譬如为计算机、平板电脑、具有触摸功能的手机等等,所述社交平台的用户影响力估算装置与上文实施例中的社交平台的用户影响力估算方法属于同一构思,在所述社交平台的用户影响力估算装置上可以运行所述社交平台的用户影响力估算方法实施例中提供的任一方法,其具体实现过程详见所述社交平台的用户影响力估算方法实施例,此处不再赘述。
需要说明的是,对本发明所述社交平台的用户影响力估算方法而言,本领域普通测试人员可以理解实现本发明实施例所述社交平台的用户影响力估算方法的全部或部分流程,是可以通过计算机程序来控制相关的硬件来完成,所述计算机程序可存储于一计算机可读取存储介质中,如存储在终端的存储器中,并被该终端内的至少一个处理器执行,在执行过程中可包括如所述社交平台的用户影响力估算方法的实施例的流程。其中,所述的存储介质可为磁碟、光盘、只读存储器(ROM,Read Only Memory)、 随机存取记忆体(RAM,Random Access Memory)等。
对本发明实施例的所述社交平台的用户影响力估算装置而言,其各功能模块可以集成在一个处理芯片中,也可以是各个模块单独物理存在,也可以两个或两个以上模块集成在一个模块中。上述集成的模块既可以采用硬件的形式实现,也可以采用软件功能模块的形式实现。所述集成的模块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,所述存储介质譬如为只读存储器,磁盘或光盘等。
以上对本发明实施例所提供的一种社交平台的用户影响力估算方法及装置进行了详细介绍,本文中应用了具体个例对本发明的原理及实施方式进行了阐述,以上实施例的说明只是用于帮助理解本发明的方法及其核心思想;同时,对于本领域的技术人员,依据本发明的思想,在具体实施方式及应用范围上均会有改变之处,综上所述,本说明书内容不应理解为对本发明的限制。

Claims (13)

  1. 一种社交平台的用户影响力估算方法,包括:
    获取社交平台上的用户行为数据;
    根据所述用户行为数据,确定用户两两之间的影响力转移关系;
    基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;
    根据所述影响力排名确定各用户的影响力。
  2. 根据权利要求1所述的社交平台的用户影响力估算方法,其中,所述根据所述用户行为数据,确定用户两两之间的影响力转移关系,包括:
    根据所述用户行为数据,生成影响力转移矩阵;
    根据所述影响力转移矩阵,确定用户两两之间的影响力转移关系。
  3. 根据权利要求2所述的社交平台的用户影响力估算方法,其中,所述根据所述用户行为数据,生成影响力转移矩阵,包括:
    基于所述用户行为数据确定第一互动信息和第二互动信息,所述第一互动信息为社交平台上用户对于好友个人发布的消息的互动信息,所述第二互动信息为社交平台上用户对于广告投放系统投放的广告的互动信息;
    根据所述第一互动信息和所述第二互动信息生成影响力转移矩阵。
  4. 根据权利要求1至3任一项所述的社交平台的用户影响力估算方法,其中,所述基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名,包括:
    获取用户在所述社交平台上的初始化影响力排名和历史影响力排名,所述历史影响力排名为上一时刻用户在所述社交平台上的影响力排名;
    通过预设网页排名算法,基于所述影响力转移关系、所述初始化影响力排名以及所述历史影响力排名,估算当前影响力排名;所述当前影响力排名为当前时刻用户在所述社交平台上的影响力排名。
  5. 根据权利要求4所述的社交平台的用户影响力估算方法,其中,所 述估算当前影响力排名之后,还包括:
    根据所述历史影响力排名以及所述当前影响力排名,估算用户最终影响力排名;
    将所述最终影响力排名确定为用户在所述社交平台上的影响力排名。
  6. 根据权利要求5所述的社交平台的用户影响力估算方法,其中,所述根据所述历史影响力排名以及所述当前影响力排名,估算用户最终影响力排名,包括:
    若所述历史影响力排名以及所述当前影响力的差别满足预设收敛条件,则将当前影响力排名确定为最终影响力排名的估算结果。
  7. 一种社交平台的用户影响力估算装置,包括:
    获取单元,配置为获取社交平台上的用户行为数据;
    第一确定单元,配置为根据所述用户行为数据,确定用户两两之间的影响力转移关系;
    估算单元,配置为基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;
    第二确定单元,配置为根据所述影响力排名确定各用户的影响力。
  8. 根据权利要求7所述的社交平台的用户影响力估算装置,其中,所述第一确定单元,包括:
    矩阵生成子单元,配置为根据所述用户行为数据,生成影响力转移矩阵;
    第一确定子单元,配置为根据所述影响力转移矩阵,确定用户两两之间的影响力转移关系。
  9. 根据权利要求8所述的社交平台的用户影响力估算装置,其中,所述矩阵生成子单元,配置为基于所述用户行为数据确定第一互动信息和第二互动信息,所述第一互动信息为社交平台上用户对于好友个人发布的消 息的互动信息,所述第二互动信息为社交平台上用户对于广告投放系统投放的广告的互动信息;根据所述第一互动信息和所述第二互动信息生成影响力转移矩阵。
  10. 根据权利要求7至9任一项所述的社交平台的用户影响力估算装置,其中,所述估算单元包括:
    获取子单元,配置为获取用户在所述社交平台上的初始化影响力排名和历史影响力排名,所述历史影响力排名为上一时刻用户在所述社交平台上的影响力排名;
    估算子单元,配置为通过预设网页排名算法,基于所述影响力转移关系、所述初始化影响力排名以及所述历史影响力排名,估算当前影响力排名,所述当前影响力排名为当前时刻用户在所述社交平台上的影响力排名。
  11. 根据权利要求10所述的社交平台的用户影响力估算装置,其中,所述估算子单元,配置为根据所述历史影响力排名以及所述当前影响力排名,估算用户最终影响力排名,将所述最终影响力排名确定为用户在所述社交平台上的影响力排名。
  12. 根据权利要求11所述的社交平台的用户影响力估算装置,其中,所述估算子单元,配置为若所述历史影响力排名以及所述当前影响力的差别满足预设收敛条件,则将当前影响力排名确定为最终影响力排名的估算结果。
  13. 一种计算机存储介质,该存储介质包括一组指令,当执行所述指令时,引起至少一个处理器执行包括以下的操作:
    获取社交平台上的用户行为数据;
    根据所述用户行为数据,确定用户两两之间的影响力转移关系;
    基于所述影响力转移关系,估算用户在所述社交平台上的影响力排名;
    根据所述影响力排名确定各用户的影响力。
PCT/CN2017/070503 2016-01-07 2017-01-06 社交平台的用户影响力估算方法、装置及计算机存储介质 WO2017118426A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/933,891 US20180211335A1 (en) 2016-01-07 2018-03-23 Method and apparatus for estimating user influence on social platform

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610009657.3 2016-01-07
CN201610009657.3A CN106952166B (zh) 2016-01-07 2016-01-07 一种社交平台的用户影响力估算方法及装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/933,891 Continuation US20180211335A1 (en) 2016-01-07 2018-03-23 Method and apparatus for estimating user influence on social platform

Publications (1)

Publication Number Publication Date
WO2017118426A1 true WO2017118426A1 (zh) 2017-07-13

Family

ID=59273264

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/070503 WO2017118426A1 (zh) 2016-01-07 2017-01-06 社交平台的用户影响力估算方法、装置及计算机存储介质

Country Status (3)

Country Link
US (1) US20180211335A1 (zh)
CN (1) CN106952166B (zh)
WO (1) WO2017118426A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110942345A (zh) * 2019-11-25 2020-03-31 北京三快在线科技有限公司 种子用户的选取方法、装置、设备及存储介质

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107894998B (zh) * 2017-10-24 2019-04-26 迅雷计算机(深圳)有限公司 视频推荐方法及装置
CN109948037A (zh) * 2017-11-24 2019-06-28 腾讯科技(深圳)有限公司 信息投放方法、装置及存储设备
CN110020154A (zh) * 2017-12-04 2019-07-16 北京京东尚科信息技术有限公司 用于确定用户影响力的方法及装置
CN109242710B (zh) * 2018-08-16 2022-03-11 北京交通大学 社交网络节点影响力排序方法及系统
CN110071838A (zh) * 2019-04-29 2019-07-30 中国石油大学(华东) 一种基于链路预测的Android智能手机节能方法
WO2021028988A1 (ja) * 2019-08-09 2021-02-18 日本電気株式会社 ユーザ位置計算装置及び方法、活動範囲推定装置及び方法、並びにコンピュータ可読媒体
CN110717823B (zh) * 2019-09-29 2022-08-02 支付宝(杭州)信息技术有限公司 一种信贷逾期风险识别方法和系统
CN111092804B (zh) * 2019-12-05 2022-03-08 腾讯科技(深圳)有限公司 信息推荐方法、信息推荐装置、电子设备和存储介质
CN111950865A (zh) * 2020-07-24 2020-11-17 黑龙江阳光惠远知识产权运营有限公司 一种识别核心专利的评价方法
CN113781250A (zh) * 2020-09-14 2021-12-10 北京沃东天骏信息技术有限公司 社交媒体信息传播评估方法和装置
CN112732780B (zh) * 2020-12-30 2024-04-02 深圳市网联安瑞网络科技有限公司 人物网络活跃度计算方法、系统、处理终端、计算机设备
CN115865708B (zh) * 2023-02-14 2023-05-02 国家计算机网络与信息安全管理中心江苏分中心 一种基于sir-d模型的复杂社交网络信息处置方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102663101A (zh) * 2012-04-13 2012-09-12 北京交通大学 一种基于新浪微博的用户等级排序算法
CN103886105A (zh) * 2014-04-11 2014-06-25 北京工业大学 一种基于社交网络用户行为的用户影响力分析方法
CN104035987A (zh) * 2014-05-30 2014-09-10 南京邮电大学 一种微博网络用户影响力排名方法
US8977617B1 (en) * 2012-10-31 2015-03-10 Google Inc. Computing social influence scores for users
CN105005918A (zh) * 2015-07-24 2015-10-28 金鹃传媒科技股份有限公司 一种基于用户行为数据和潜在用户影响力分析的在线广告推送方法及其推送评估方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8250069B2 (en) * 2008-10-14 2012-08-21 University Of Washington Green's function formulations for pagerank algorithm using helmholtz wave equation representations of internet interactions
CN102270204A (zh) * 2010-06-02 2011-12-07 上海佳艾商务信息咨询有限公司 一种基于矩阵分解对在线论坛用户影响力进行计算的方法
JP5727421B2 (ja) * 2012-07-20 2015-06-03 日本電信電話株式会社 関連ノード検索装置、関連ノード検索方法、及びプログラム
CN103678669B (zh) * 2013-12-25 2017-02-08 福州大学 一种社交网络中的社区影响力评估系统及方法
CN104268648B (zh) * 2014-09-28 2018-01-12 福州大学 融合用户多种交互信息和用户主题信息的用户排名系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102663101A (zh) * 2012-04-13 2012-09-12 北京交通大学 一种基于新浪微博的用户等级排序算法
US8977617B1 (en) * 2012-10-31 2015-03-10 Google Inc. Computing social influence scores for users
CN103886105A (zh) * 2014-04-11 2014-06-25 北京工业大学 一种基于社交网络用户行为的用户影响力分析方法
CN104035987A (zh) * 2014-05-30 2014-09-10 南京邮电大学 一种微博网络用户影响力排名方法
CN105005918A (zh) * 2015-07-24 2015-10-28 金鹃传媒科技股份有限公司 一种基于用户行为数据和潜在用户影响力分析的在线广告推送方法及其推送评估方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110942345A (zh) * 2019-11-25 2020-03-31 北京三快在线科技有限公司 种子用户的选取方法、装置、设备及存储介质
CN110942345B (zh) * 2019-11-25 2022-02-15 北京三快在线科技有限公司 种子用户的选取方法、装置、设备及存储介质

Also Published As

Publication number Publication date
US20180211335A1 (en) 2018-07-26
CN106952166A (zh) 2017-07-14
CN106952166B (zh) 2020-11-03

Similar Documents

Publication Publication Date Title
WO2017118426A1 (zh) 社交平台的用户影响力估算方法、装置及计算机存储介质
US10504150B2 (en) Location retargeting system for online advertising
US11100534B2 (en) Behavioral retargeting system and method for cookie-disabled devices
US9336282B2 (en) Systems and methods for identifying and analyzing internet users
JP6145576B2 (ja) オンライン・ソーシャル・ネットワークにおける大規模ページ推薦
US11587143B2 (en) Neural contextual bandit based computational recommendation method and apparatus
US9483580B2 (en) Estimation of closeness of topics based on graph analytics
US20150134402A1 (en) System and method for network-oblivious community detection
US20170171294A1 (en) Computerized System And Method For Determining And Communicating Media Content To A User Based On A Physical Location Of The User
US20140122697A1 (en) Providing content to linked devices associated with a user
US20150242447A1 (en) Identifying effective crowdsource contributors and high quality contributions
US20170140397A1 (en) Measuring influence propagation within networks
US11263664B2 (en) Computerized system and method for augmenting search terms for increased efficiency and effectiveness in identifying content
JP6059314B1 (ja) 推定装置、推定方法及び推定プログラム
CN108401005B (zh) 一种表情推荐方法和装置
US20130275223A1 (en) Future ad targeting
WO2015196561A1 (zh) 一种用于提供发布信息的方法与设备
US10922722B2 (en) System and method for contextual video advertisement serving in guaranteed display advertising
US20190066231A1 (en) Methods and systems for providing content to a user of a relationship network
JP2017059255A (ja) 推定装置、推定方法及び推定プログラム
CN110020129B (zh) 点击率校正方法、预估方法、装置、计算设备及存储介质
US11973841B2 (en) System and method for user model based on app behavior
US11977563B2 (en) Cross-entity categorization of heterogenous data
JP7146037B1 (ja) 情報処理装置、情報処理方法、及び情報処理プログラム
CN117980924A (zh) 隐私保护机器学习扩展模型

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: 17735864

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: 17735864

Country of ref document: EP

Kind code of ref document: A1