Network Public Opinion Analysis Based on LDA Model

In order to effectively mine hidden research topics and potential evolution patterns from massive network public opinion data, this paper proposes a public opinion monitoring system based on LDA. The scheme mainly performs text crawling and topic extraction of network entities through self-programming by Python, and realizes fine-grained emotion analysis and topic mining combined with TF-IDF feature words and LDA model. Then, cluster analysis is made on the content and intensity of each stage of public opinion development, and the construction of topic model and topic confusion are extracted for further public opinion prediction. Finally, the LDA topic model is applied to the application of public opinion system, which realizes the intelligent monitoring, analysis and prediction function of public opinion, and it can quickly respond to the update and change of network public opinion.