Research On The Sentiment Analysis Method Of Online Public Opinion Based On Picture Fuzzy Set

With the rapid development of internet technology and social platforms,the public is more and more willing to express their opinions on the internet.These comments carry the public’s emotional tendencies,and the analysis of these emotional tendencies has a very important role for enterprises to formulate strategies and monitor public opinion.Therefore,it is of theoretical and practical significance to continuously propose more effective sentiment analysis methods,which can not only promote the development of the Chinese NLP field but also better serve the practical problems such as market research,user analysis,and online opinion discovery and early warning in practice.The main research contents of the article are as follows.(1)Considering that current text representation methods ignore the sentiment diversity of texts,this paper builds a text sentiment representation model called Word2 PFS based on the picture fuzzy sets(PFS).The method combines the emotional dictionaries and PMI algorithm ideas to explore the positive,neutral,and negative sentiments embedded in texts.The sentiment information of the text is expressed in the form of PFS.Compared to other word vector models,Word2 PFS is a better model for sentiment classification problems.(2)The majority of current sentiment classification methods rely on corpus training,and the sentiment polarity of a comment text cannot be simply classified as positive or negative.This characteristic can also make manual sentiment labeling more difficult.Therefore,we consider the study of clustering algorithms for text sentiment clustering.The new fuzzy clustering algorithm MKPFCM-GWO is proposed by optimizing and improving the fuzzy C-mean clustering algorithm based on the theory of multiple kernels and group intelligent optimization algorithms.(3)The text sentiment analysis method proposed in the paper is applied to a practical problem.We obtain the comments on the COVID-19 on microblogs,and the sentiment analysis of the comment text is carried out by Word2 PFS and MKPFCM-GWO models.In addition,a series of text mining techniques are used to explore the changes in public sentiment and the hot topics over time,and to provide support for the formulation of public opinion response plans.The main contributions of this paper are two: firstly,it is the first time to introduce the picture fuzzy set into the word vector method and construct the Word2 PFS word vector method.Secondly,we consider the application of picture fuzzy clustering algorithm in text sentiment analysis and combines the Word2 PFS method to conduct the research of text sentiment analysis of microblog public opinion.