Application Research On Network Support Degree For Government Decision Based On Machine Learning

Network public opinion has become one of the most important components of contemporary public opinion,and it is also an important basis and data source for the government to listen to public opinion,grasp the direction of public opinion and make government decisions.Meanwhile,machine learning has become one of the most effective data analysis methods used in the field of data science.Although there are many researches using machine learning to analyze emotions,there is a lack of research in the field of government decision-making,and its accuracy and reliability are not ideal.Therefore,we use the method of machine learning to analyze the emotional tendency of weibo comments related to government decision-making,so as to obtain the degree of netizens' support for a certain government decision-making and provide reference for the government.Aiming at the particularity of government decision-making micro-blog,this paper creatively constructs a dictionary of emotional analysis in the field of government decision-making,sets up an emotional scoring mechanism,and adds web spammer detection before emotional analysis,in order to improve the accuracy and reliability of judging the support degree of micro-bloggers for government decision-making.This paper tries to select all the comments below 30 government decision-making microblogs on Sina Weibo.First,we use an emotional orientation analysis method based on the emotional dictionary and emotional scoring mechanism in the field of government affairs to classify the comments,and use different classification models to train and compare their effects.Then we use an improved multi-feature-based web spammer detection method to detect and filter the comments on each micro-blog,and then use the above method to classify the emotional orientation.The experimental results show that adding web spammer detection before classification can further improve the accuracy and reliability of judging the support degree of micro-bloggers for government decision-making.