关键词: 网络舆情, 情感值, 组合模型, 预测

Abstract:

As the number of significant events of public opinion surged in recent years,it is extremely serious to keep the security of the Network public opinion and to warn early.Thus,it is important to grasp the trend of public opinion in the macro.This paper presents a network public sentiment forecast based on ARIMA and BP neural network.The emotional value of the text of "the flood incident in Yuyao" on Sina Weibo is calculated.Then the comprehensive analysis and forecast of the trend of emotional change are given.Forecast results show that relative to,method of combined forecasting is more precise than the single forecast method,indicating the application of this model in the emotion forecast of Internet public sentiment is feasible.

Key words: public opinion;emotional value;combined model;forecast