基于BP神经网络的网络舆情预警研究

Research on Early Warning of Public Opinion Based on BP Neural Network

摘要: 本文首先构建了包含4个一级指标和16个二级指标的网络舆情预警指标体系，然后利用BP神经网络建立网络舆情预警模型对网络舆情事件进行等级预警，并对该模型进行了实证分析。结果表明，舆情主题、舆情热度、舆情参与度、舆情传播态势这几个指标对预警等级的影响较大，然后利用BP神经网络进行网络舆情预警的正确率高达70%，但是预警的平均精确率为62.5%，还有提升空间。总体上来说，网络舆情预警指标体系具有可行性，且BP神经网络预警模型是有效的，可以为网络舆情事件的预警方法提供理论参考。

Abstract: In the context of social stability, this paper builds a network public opinion early warning index system that includes 4 primary indicators and 16 secondary indicators. The BP neural network is used to establish a network public opinion early warning model to give a level early warning of network public opinion events, and to carry out the model empirical analysis. The results show that public sentiment theme, public sentiment, public sentiment participation, and public opinion dissemination trends have a large impact on the level of early warning. Then, the accuracy rate of online public opinion warning using BP neural network is as high as 70%, but the average accuracy of the warning 62.5%; there is still room for improvement. In general, the network public opinion early warning index system is feasible, and the BP neural network early warning model is effective, which can provide a theoretical reference for the network public opinion early warning method.