Text Classification And Public Opinion Analysis Of "Garbage Classification"

With the continuous improvement of people’s living standards,people’s pursuit for material become rich and colorful,and various kinds of garbage in daily life are produced rapidly.Every country in the world is looking for an effective way to solve the problem of waste pollution which is suitable for its own situation.At present,China is facing a serious dilemma of "garbage siege".In the middle of 2019,Shanghai implemented the regulations on waste classification management,then 46 cities across the country launched waste classification pilot work.For a while,the topic of "garbage classification" was widely discussed by netizens.To promote the effective implementation of waste classification policies,it is not only the government that issues relevant policies to supervise and sanction,but also on the consciousness of every citizen,the correct establishment of the concept of classification and the effective promotion of the education of the general public awareness.In the process of promoting garbage classification nationwide,there is a public opinion upsurge focusing on it on the Internet.In this thesis,we use the technology of web crawler to collect the text comments about "garbage classification" in Sina Weibo and Zhihu,and manually search the opinions and opinions of netizens in other social platforms,with a total of 5000 text data.After keyword screening,we retain 2000 remaining valid texts,and then preprocess the text.The text training set is manually labeled as positive and negative.In the section to stop words,there are some unique non emotional words in the comments on the topic of "garbage classification".Therefore,this thesis has customized a disable thesaurus,which combines with the original stop words list,and improves the quality of feature extraction.Secondly,three feature extraction methods,mutual information,information gain and chi square test,are used for comparison,and the best chi square statistics and TF-IDF method of keyword feature extraction are used to build the subsequent classification model.Compared with naive Bayes,logistic regression and support vector machine,three classification algorithms are used to further mine the views and attitudes of netizens.The results show that chi square statistics is used to improve The accuracy of the classification is the best.Finally,the public opinion analysis is carried out with the classification results of the selected model.Respond to the opinions of netizens,which provides public opinion support for the follow-up progress of the publicity of garbage classification policy issues.At the same time,through the heated discussion of netizens on microblog,Zhihu and other social platforms,it provides some positive and effective opinions and suggestions for the current garbage classification policy.