Research on Cross-platform Network Public Opinion Risk Early Warning Based on LDA and TextCNN
[Purpose/significance]It is of great social significance and practical value to analyze network public opinion data on multiple social platforms,assess network public opinion risks,and conduct risk early warning research.[Method/process]The paper constructs the network public opinion risk index system firstly,and then uses the analytic hierarchy process to determine the index weight,so as to build the network public opinion risk early warning model.The empirical part uses the network public opinion data of a prefecture-level city for analysis.The paper uses LDA to cluster the data on Weibo platform firstly,and then applies the TextCNN to classify the data of other social platforms according to the clustered data,and finally employs network public opinion risk early warning model to study the public opinion on various topics.[Result/conclusion]The network public opinion risk early warning model construc-ted in this paper has certain accuracy and effectiveness.The network public opinion risk early warning model in this paper can provide information support to improve decision-making efficiency and monitoring efficiency of network public opinion risk.
network public opinionrisk early warningtopic clusteringtext classification
万方数据
维普
