Research On Personality-aware Friend Recommendation Algorithm Based On Social Network

With the popularity of social networking and gradually deepening into all aspects of people's lives,communication and interaction by network is becoming a main way of communication.Micro-blog is an online social networking platform,which has the characteristics of openness and real-time,and it is a platform based on information sharing,dissemination,and acquisition of user relationships.If you can accurately recommend micro-blog friends that they are interested in,this will improve the economic benefits of micro-blog products(accurate advertising)and improve the user experience(good friend recommendation).However,most of the existing bloggers' recommendations rely on topological structure or content features to make recommendations.In social psychology,people interact with each other in their daily lives and intrinsic personality is one of the main factors affecting people's behavior.This paper aims to improve the accuracy of micro-blog friend recommendation by considering the influence of user personality on user behavior in micro-blog friend recommendation.In view of the above problems,the main contents of this research work include the following points:1.The Big Five personality feature combined with the basic characteristics of Weibo users as the followee recommended similarity rating standard.First,we calculate Weibo-specific features,including location information,micro-blog text,social information,interactive information and other comprehensive similarity scores.Then we calculate the big five personality types of micro-blog users.Finally,synthesizing two scores can recommend potential followee for users.2.Based on the analysis of the basic features of micro-blog users,this paper proposes a PSER model based on the Big Five personality,micro-blog semantic similarity and micro-blog emotional analysis.Because of the inherent consistency of personality,emotion and semantics.Therefore,the selection of PSER model can improve the effect of improving the quality of recommendation.Finally,in order to verify the validity of the proposed model and algorithm,we will carry out relevant experiments and analysis on the real data set.The first experimental result shows that the recommendation accuracy rate has improved significantly after adding the Big Five personality feature to the basic characteristics of micro-blog users.The second experimental result shows that the PSER model that combines the Big Five personality,micro-blog text semantic similarity and sentiment analysis has a very significant improvement in the recommendation accuracy rate and recall rate.In the end,compared with the traditional recommendation method,the experiment shows that joining the personality-aware recommendation method of Big Five personality has outstanding effect and practical significance on the blogger's recommendation for the user's behavior and the user's personality.