Research And Implementation Of Network Public Opinion Analysis Technology Based On Machine Learning

With the rapid development of computer science and technology,the number of netizens is constantly increasing,and the "cohesive" of people's lives and the Internet is getting higher and higher.For the acquisition of information,the expression of opinions and the venting of emotions,netizens are increasingly relying on online social platforms.The online text data contains great social value and commercial value.How to analyze the massive online commentary text data to obtain the opinions and emotional tendencies of the masses is a popular topic in current research.The main research direction of this paper is the emotional classification of texts.Through the in-depth study of related technologies,the text sentiment classification based on machine learning is realized,and the network sentiment analysis is realized from the analysis of the sentiment orientation of the text.The main work of this paper is as follows:(1)The experiments in this paper require a large amount of experimental data.Currently,the open source data is too small to apply.In this paper,a web spider is designed and implemented,and a large amount of comment text data is captured from Sina Weibo.The collected data is collated,manually labeled,and pre-processed to provide experimental data for subsequent work.(2)Based on the custom text data set and reference text data set,the support vector machine(SVM)classification model is used for experimental test,and the quality of the custom text data set is verified through the comparison of model performanc.(3)Through the intensive study of the relevant knowledge of capsule network image identification,the text sentiment classification model of CapsNet is designed.SVM classifier and CapsNet model are trained by the customized text data set and the reference text data set respectively.And using the CapsNet model to implement network public opinion analysis.The performance of the finally trained CapsNet text emotion classification model and SVM classification model is tested through the test data set.The experimental results show that the performance of the CapsNet is better than that of the SVM classifier.