Research On User Interest Mining Based On Social Network Content Analysis

As the development of Internet technology,we are moving into a fast-tempo society with our work and life,so the traditional social media has can not satisfy our needs,more and more people choose the online social network.According to China Internet Network Information Center(CNNIC),In the existing online social platform,integrated social applications have a lot of users,often as people's daily information sharing and dissemination of the platform.For the weak social network,the user has a high degree of freedom,the use of platform attention hot news,concerned about the content of interest,access to knowledge and sharing point of view.Therefore,the analysis of user interests for social platforms is very important.According to psychological research,user interest has the characteristics of migration and forgetting,which is divided into long-term interest and short-term interest on the basis of attention time.This paper combines the interest attention probability and time to define the sequence of user interest.Considering the limitation of user's energy and the divergence of interest determined by subjective intention,this paper propose a limited resources artificial immune system user interest mining model based on fuzzy resource allocation strategy.The users of social text preprocessing as intrusion antigen by similarity between antigen and memory cells corresponding text feature words,to determine if there has been a user interest,which gives the corresponding immune response.Then constructs the interest concern sequence,calculates the user interest probability,thus reflects the user interest change.Through comparative experiments,the user interest mining model based on social content can give users the interest degree of different time,and well reflect the changes in user interest.The interests of people in social networks affect each other.Because of different purposes,users will be divided into information seeking users and information sharing user.Especially for information seeking users,their related social contents are less,based on the user's text content user interest model has some problems.This paper constructs a user interest mining model based on social relations,and further optimizes the user interest vector.According to the quality of user fans using Page Rank algorithm for user importance ranking,and use the user importance ranking for user classification.Analysis of the factors affecting users in social networks,and the weight of these factors are given by gradient ascent algorithm,and then gives the social impact value of the relevant user.Finally,for different types of users,give their comprehensive social relationships and social content user interests.Through comparative experiments,on the basis of social content,consider the impact of social relationships on user interest,get more comprehensive user interests.