Font Size: a A A

Design Of Potential Social Relationship Discovery Schemes For Users Based On Topic Models

Posted on:2019-02-26Degree:MasterType:Thesis
Country:ChinaCandidate:H K LeFull Text:PDF
GTID:2358330548457619Subject:Applied statistics
Abstract/Summary:PDF Full Text Request
Internet has been developing rapidly in recent years,information dissemination into the era of new media,the various types of social media platforms,live platform,electric business platform,new media platforms such as news sites have sprung up,the platform of registered users is also growing rapidly.Although the users of these platforms have a large number of users,due to the low social properties of users on the platform,the platform has more choice,which leads to low stickiness,low loyalty and easy loss of users and platforms.In order to find out the potential social relationship between users on the new media platform,increase the number of users' friends and improve the user's social behavior strength,thus further forming the user community to improve user loyalty.This paper systematically studies the discovery methods of potential social relationships among new media users.Because of the new media platform for user traffic,so often simplified steps when user registration allows the user to register a convenient,this leads to a platform to grasp user's personal data,lack of social attributes.Some ways to find potential social relationships based on user data similarity and users' existing social relationships are difficult to use.But on the other hand,due to the development of the intelligent equipment and large data application is more and more widely,new media platforms and tend to have very good interaction between the user,the user will be after reading the information on the platform point forward,reviews,and post operation,these data will be recorded,and these data is reflected the user interest preferences.Therefore,this article,through the establishment of the interest preference recommends theoretical framework,and apply the LSA(Latent Semantic Analysis)and LSTM(Long Short-Term Memory)implement,established a set of new media users against potential social relations found the solution.Finally,this paper takes a football news App as an example to analyze the application of this scheme in real business situations.
Keywords/Search Tags:LDA, LSTM, Potential Social Relationship, Emotion Classification
PDF Full Text Request
Related items