Investor sentiment and stock market volatility: Exploring the relationship using sentiment analysis of stock bar comments

Emotional comments on social media platforms, such as stock review websites, can consequently affect the behavior of stock investors and the volatility of stock prices. Current approaches to stock price prediction frequently neglect the influential role of investor sentiment or do not investigate the interconnection between them. Furthermore, a sentiment dictionary specifically tailored to the stock market remains notably absent. The present study begins with standardization and preprocessing operations on both the stock price index and stock commentary big data. The investor sentiment indicator is subsequently constructed using sentiment dictionaries and comments from stock forums. Furthermore, the correlation between two sets of big data is investigated utilizing the quantified investor sentiment indicator acquired from comment data and the big data of the stock price index. This study provides a direct basis for subsequent applied research on stock price prediction engineering, which is essential for informed decision-making by investors and managers. The text data are quantified and converted into an investor sentiment indicator using the sentiment analysis method based on the sentiment dictionary. The correlation between the investor sentiment indicator and the price fluctuation of the CSI 300 index is studied using the SPSS correlation analysis method and the deep learning method. The experimental results show a positive correlation between investor sentiment based on stock bar comments and the closing price and the return rate of the CSI 300 index, with mutual influence. The impact of the investor sentiment indicator on the stock closing price lags, and the correlation gradually decreases with the time between the two variables. In addition, the investor sentiment indicator positively correlates with the stock market in different stages, with a greater impact during the rising stage than in the falling stage. • Constructed a stock-market-specific sentiment dictionary from forum comments. • Quantified investor sentiment and CSI 300 index data after preprocessing. • Found positive, lagged correlation between sentiment and index returns. • Sentiment impacts stronger in rising than in falling market stages.