Research On Micro-blog Users' Behavior Based Sentiment And Interest

With the growing diversity of social networks,the growing size of the user,the user's social life and the virtual life are more closely linked.The research on the behavior prediction of micro-blog users' praise,comment and forwarding is of great significance and value for the research of personalized recommendation,network real-time monitoring and guiding public opinion.This paper mainly analyzes the factors that affect the behavior of microblogging users from three aspects: user attribute,user interest and user emotion.Based on the results of statistical analysis,the key features are selected to construct the forwarding,comments,and likes behavior prediction model.1.In the aspect of user's emotion,we use the short text analysis method to analyze the content of microblogging text,build the emotional dictionary and establish the emotional classification model,and realize the emotional classification of microblogging text.In the user interest model,the user labels features and micro-blog hidden topic feature are fused to establish Real time accurate extraction model a interest model.2.In the micro-blog user behavior prediction model,the characteristics of the emotion and interest are fused with the user's attributes,and the prediction model is trained in the naive Bias,K nearest neighbor and support vector machine classifier.The results show that the average accuracy rate of emotion classification can reach more than 85%,and the average accuracy rate of interest classification is more than 84%,which verifies the validity of emotion model and interest model.Moreover,the behavior prediction experiment also compares the emotion and interest characteristics in the model.The results show that the average accuracy of forwarding behavior can reach 82.56% in the prediction,the average prediction accuracy of behavior in the comments can reach 84.59%,the prediction average accuracy of likes behavior rate can reach 79.35%,which indicates the effectiveness of user interest and emotion characteristics in the promotion of microblogging user behavior prediction.