Research On Sentiment Tendency Of Network Public Opinion Based On Machine Learning

In recent years,people can conveniently express their emotions and opinions through microblogs,forums and other carriers due to the development of mobile Internet technology.The information published by users often contains a certain degree of emotional tendency and opinion characteristics.In the fields of public opinion analysis and monitoring,it is significant to study these characteristics through sentiment orientation analysis.This paper takes the forum text data as the research object and conducts research using the method of emotional tendency analysis.The specific contents are as follows:Firstly,the paper introduces the background and significance of the public opinion tendency research on the forum data,and expounds the research status of sentiment analysis technology.At the same time,this paper explains the process and related technology of sentiment orientation analysis,including text collection technology,text preprocessing technology,text representation technology,performance evaluation indicators and so on.Then,the paper uses the naive Bayesian algorithm and the dictionary method to analyze the political tendency of the forum text.After comparing the applicability of the algorithm,the dictionary method with better effect is finally selected.The algorithm is optimized four times by expanding the segmentation database,expanding the polar vocabulary.The final average accuracy rate reached 87%,and the average recall rate reached 81%,which can better reflect the emotional tendency of the text.