Computer Intelligent Prediction Method for Hot Event by Long Short-Term Memory Model

At present, with the development of the Internet, people can timely post their cognitions, attitudes and emotions about hot events, which constitute the network public opinions. Through the analysis of the network public opinions and sentiments, the evolution of hot events can be judged and further provided suggestion to relevant departments to deal with hot event crisis. The exiting approaches still have some shortages, such as inaccurate classification of sentimants and topics, poor predicting capability of traditional prediction models. To sovle those problems, we propose a novel approach, which is based on sentiment analysis, text clustering and LSTM neural network model. First, we extracted emotional tendencies to classify text comments into positive and negative stentiments with SnowNLP, a Chinese natural language processing tool. Second, based on the Single-Pass clustering algorithm, we implemented text cluster analysis to explore hot topics. At last, we predicted the public emtional trends and the topic attention trends with a Long-Short-Term Memory model(LSTM). The experiments show that our approach has a high accuracy of prediction about the evolution trend of hot events.