User Behaviours Prediction in Microblog Based on Human Knowledge Cognitive Process

As Social Media is opening up, users usually write blogs to report the real-world events and make comments on the Web. When a blog emerges on the web to reflect a hot event, a large volume of comments follows this blog with different content and emotion. For the author of a blog, how to select reprehensive comments from so large number of comments to reply is a challenging issue. To solve this issue, we propose a user behaviour prediction method based on the cognitive dissonance theory. The method constructs user prior knowledge network as user profile, which consists of content and emotion by history microblogs of the user. By comparing the user´s current microblog with the comments follow it, we calculate dissonance values of the content and emotion respectively to predict user behaviour. The behaviour is about which comments that the user are preferential to reply. Our method can predict which comments the user preferentially to reply. The experimental results show that our method can previously predict which comment to reply with higher accuracy.