Research On Personality Analysis Methods Of Social Network Users Based On Topics

Personality is a personality characteristic unique to individuals.Its formation is related to many factors,and is relatively stable and will accompany people's life,affecting people's behavior.The emergence of social networking platforms has created conditions for researchers to study user personality traits,and user network behavior data can accurately reflect user personality traits.The growing convergence between social sciences and computer science has led researchers to develop automated methods to extract and analyze user data footprints to analyze personality traits.This article uses Sina Weibo as a social network platform to study the personality categories of users,and uses the topics in user search hotspots as research objects to analyze in depth the personality categories reflected by users'attitudes and comments on different events.The main work of this article includes:(1)A social personality analysis method based on CI-SVM model was established.This part studies the influence of group intelligence generated by a small group of users participating in the same topic on the analysis of individual personality traits in the group under the network environment,and then creates a eigenvalue calculation expression combining with group intelligence,and then constructs CI-SVM model for social network user personality analysis.Personality analysis of social network users.The experimental results show that the constructed CI-SVM model is valuable for personality analysis of social network users.This helps a lot in social network user personality analysis.(2)Build a RAkEL-PA model based on social networks.In order to fully consider the correlation between the five personality categories,the multi-label integration method RAkEL was improved and the RAkEL-PA model was constructed.The experiments extracted topic-related data from Weibo users to extract characteristics related to personality categories.The experiments were performed on two RAkEL_d-PA and RAkEL_o-PA submodels.The experimental results show that the RAkEL-PA model based on social networks improves the accuracy of user personality category recognition.At the same time,the model is compared with the four commonly used multi-label learning methods of BR,CC,MLkNN and LP.