| With the development of financial markets,behavioral finance has received increasing attention and has been used to analyze many phenomena in the real economic market that are difficult to fully explain by efficient market hypotheses.Due to the rapid popularization and development of the Internet,it is not only an important channel for information dissemination,sharing,exchange,and interaction,but also an important venue for public discussion.Here,people can easily grasp various valuable news,thereby better understanding market changes and grasping market trends.Investors in the real market are influenced by the public opinion of the online stock market,which can bring personal emotions into the decision-making process and affect the decisions of themselves and surrounding investors.Processing and analyzing text information of online stock market public opinion through computer technology,mining valuable information in online stock market public opinion,and quantitatively analyzing its relationship with market trends has become an important topic in the field of behavioral finance research.This article summarizes domestic and foreign literature,conducts research and analysis on relevant theories and technologies,and combines natural language processing technology and machine learning algorithms to analyze online stock market public opinion data.Key phrases representing key information are extracted,and a regression model based on machine learning is used to analyze their relationship with market indicators,thereby guiding investors in investment decisions.The main research work and results of this article are as follows:(1)Text processing and quantitative analysis methodsThrough research on text processing and analysis methods,using online stock market public opinion data,the text is cleaned,segmented,and quantified to form a basic online stock market public opinion corpus.The VSM model is used to quantify the text and provide basic data for keyword extraction.(2)Keyword extraction based on semantic and statistical weightsThrough the research on keyword extraction technology,firstly,the basic sentiment dictionary and Word2Vec method are used to achieve semantic based domain keyword extension.Then,the TF-IDF method is used to calculate the statistical weight of keywords,and combining semantic and statistical weights,keyword extension and extraction are achieved,providing a foundation for quantitative analysis of public opinion and market indicators.(3)Research on the Application of Public Opinion Data Based on Machine LearningThe keyword extraction algorithm proposed in this article demonstrates the correlation between online stock market public opinion and market indicators.At the same time,different keyword extraction algorithms and three regression models,linear regression,BP network regression,and SVR regression,were combined to further analyze the quantitative relationship between online stock market public opinion and market indicators.Combined with the results of empirical research,it was demonstrated that online stock market public opinion can analyze and predict market indicators.Meanwhile,compared to the extraction algorithm using independent semantics or statistical weights,the keyword extraction algorithm proposed in this article can more effectively and accurately describe the public opinion of the online stock market.The keyword extraction technology and regression analysis method based on machine learning proposed in this article have certain theoretical and practical guidance significance for investors and researchers.Based on relevant theories and technologies,as well as empirical evidence,the results show that the keyword extraction method combining semantics and statistical weights can effectively extract key information of public opinion in the online stock market.Based on machine learning regression models,the extracted keywords of online stock market public opinion can be used to quantitatively analyze the relationship between online stock market public opinion and market indicators,and to analyze and predict market trends.This has certain theoretical and practical guiding significance for investors and researchers. |