User Profile Construction And Its Application Based On Online Social Network

With the development of social networks,more and more data are generated by users on different platforms,which have laid a foundation for user profile construction.At present,personalized recommendation systems are getting more and more important in the of marketing and e-commerce fields.While user profile construction is the key to the success of such recommendation systems.Therefore,the study on user profile construction in online social network has practical application value.User basic attributes and interests are important components of user profile.In this paper,we first predict users' basic attributes based on user generated content from online social network.Then,a mobile terminal microblog reading system is developed in order to collect gestures of users.This system can also excavate the interest of users and construct their interest profile.Based on the basic attributes and interest profile,we build the user's profile.Finally,application of the user's profile in the user's interest recommendation is studied.This paper uses the user generated content in the social network to predict the basic attributes of the user,including the gender,age and region information.Based on the ten-fold cross validation method,after finding the optimal parameters of the topic model and vector space model in the training set,user modeling and training classifier are trained to predict the basic attributes of users in the test set.The experimental results show that the generated content of different sex and age of users in the Sina microblog in the differences in the subject,using the topic model achieved the best prediction accuracy;in the user's local prediction task,user generated content between different regions without obvious difference,prediction effect was significantly lower than the topic model the prediction effect of vector space model.Secondly,in order to study the interest of users by using gestures on mobile devices,the feasibility of building a user's interest profile is building This paper develops a mobile terminal microblog reading system.It collects microblog's content corresponding to user generated gesture and gesture based on this system,so as to tap user interest and finally build user interest profile.Experimental results show that user gesture behavior can reflect user interest in browsing process,and user interest profile construction based on gesture behavior can achieve better results.Finally,this paper constructs a complete user profile with the user's basic attributes and interest profile.In order to study the effect of user profile in the practical application,this article is based on the user's profile to recommend the microblog to the user.The experimental results show that microblog recommendation based on user profile can achieve better results,and the test users have higher satisfaction with the recommended microblog.Overall,this paper investigate user profile construction and conduct related applications of in online social network.User behavior and corresponding content on the online social network are mainly used to build user profile and do interest recommendation.Findings in this paper show a great value and significance to the research of personalized information organization and information recommendation.