Quantitative Measurement of Investor Sentiment Indicators and Theoretical Research on Stock Market Return Prediction

Abstract

This paper, from the perspective of behavioral finance, explores the quantitative measurement methods of investor sentiment indicators and their impact on stock market returns. First, define investor sentiment and summarize its mechanism of action, and explore how sentiment factors affect the pricing efficiency of the capital market. In terms of measurement methods, this paper constructs a comprehensive investor sentiment index from the perspectives of direct questionnaires, indirect market data, and text mining, and uses principal component analysis to extract the general factors in the sentiment factors, thereby overcoming the shortcomings brought by a single indicator. An empirical analysis of China's A-share market from 2010 to 2023 shows that there is A significant non-linear relationship between investor sentiment and stock market returns: sentiment is positively correlated with returns in the short term, indicating price deviations caused by sentiment. In the long term, there is a negative correlation, indicating a rational return to the market. Further predictive models showed that VAR models with sentiment variables improved yield predictions by 15% to 20% compared to traditional financial models, especially when market volatility was high. At the same time, the sentiment of different investors has a different effect on the prediction of yields. The sentiment of institutional investors has a better predictive effect on long-term yields, while the sentiment of individual investors has a better explanatory power for short-term market volatility. This paper provides empirical support for enriching and developing behavioral finance theory, and also offers useful references for investors' investment decisions and risk control.