基于股吧评论的投资者情绪与股市波动研究——以九安医疗为例

Research on Investor Sentiment and Stock Market Volatility Based on Stock Bar Comments—Taking Jiu’an Medical as an Example

摘要: 以新冠疫情爆发为研究背景，针对东方财富股吧中大量能反映股吧舆情的文本数据，本文选取九安医疗2022年1月15日到2022年10月15日的股吧评论数据，对比多种情感分析模型，选用Bert + Bi-LSTM模型对投资者情绪进行分类。样本期内积极情感占比为41%，消极情感占比为59%，利用构建的投资者情感指数与股价涨跌幅进行多元线性回归模型，揭示股吧中投资者情绪变动与股市波动有关系，并显示短期内有正向预测效果，长期内有负向预测效果。但本文实证的样本期较短，选取的股吧数据平台较单一，可能对实验结果有一定影响。

Abstract: Taking the outbreak of the COVID-19 as the research background, and aiming at a large number of text data in the Oriental Fortune Stock Bar that can reflect the public opinion of the stock bar, this paper selects the review data of the stock bar of Jiu’an Medical from January 15, 2022 to October 15, 2022, compares a variety of emotion analysis models, and selects Bert + Bi-LSTM model to classify investor sentiment. During the sample period, the proportion of positive emotions was 41%, and the proportion of negative emotions was 59%. Using the constructed investor sentiment index and the stock price rise and fall, we conducted a multiple linear regression model to reveal the rela-tionship between investor sentiment changes in the stock bar and stock market volatility, and showed that there was a positive prediction effect in the short term and a negative prediction effect in the long term. However, the sample period of this empirical study is short, and the stock bar data platform selected is single, which may have a certain impact on the experimental results.