Data Trend Analysis In The Application Of Public Opinion Monitoring System Research

With the rise of big data in network era,many of public opinion data emerge in endlessly with full of emotional color,They also drives the various views in the different public opinion direction.So the tendency of public opinion data analysis is getting more and more attention by many scholars and government officials.Public opinion always like critical text content,data tendency is appraise degree of text content which is good or bad.In this paper,We use the network comments data to analyze the tendency.With the introduction of language processing,topic extraction,machine learning and other related knowledge.We tend to achieve the goal of improving tendency classification accuracy.The main research work is as follows:1.The influencing factors of network public opinion tendency analysis.In this paper,the feature selection of public opinion analysis and theme extraction skill was studied,We found out how it influenced on tendency classification results and used the combined classifier model as the main classification method.We analyzed the different people's view had differences influence on tendency classification results,combined with the present stage of research model,optimized the relevant stage of the algorithm.2.Research on topic extraction mode.We analyzed the forming process of the subject view,realized the importance of specific theme views in the overall tendency analysis process,To view the formation process of combined with theme LDA model,and verified the LDA model for final classification effect.Process,The analysis of the tendency of phase of feature extraction algorithms joined the word frequency impact factor is optimized,and the optimized algorithm for validation,found out the most efficient way of feature extraction.3.Public opinion data tendency analysis method study.View of complexity of network public opinion data,We introduced the concept of machine learning through the combination classification principle on distributed framework classifier structures and improved the accuracy of the data trend analysis results to get the final result of tendency analysis quickly.Results showed that the theme extraction model in description of the process of public opinion view was very precisely,The theme extraction model based on combining classifier could improve the tendency analysis results.After optimization of feature selection methods,the analyzed of the tendency results have improved significantly.We combined classifier to the distributed framework,We analyzed the data trend in public opinion,the results told us that the effect in speed and accuracy has been improved significantly.