Abstract: In view of the characteristics of microblog platform,a public opinion prediction method is proposed on the basis of neural networks.In this method,the post amount in unit time is taken as the quantitative index of event trend.Then,by considering the factors influencing events,the modeling is performed according to sample data,and the neural networks are employed to predict the future trend of the events outside the scope.Simulation results show that the proposed method is fast in terms of the quantization and modeling of the event trend,and it is effective in predicting the outbreak point and the post amount.

Key words: social computing, neural networks, big data, trend forecasting, microblog