Analysis of Investor Sentiment and Trading Behavior

This work focuses on analyzing the non-ferrous metal stock market by selecting specific indicators to reflect the sector's overall performance, considering the prevalent market conditions. In this paper, we preprocessed the index data and applied principal component analysis to obtain five principal components, accounting for 95.263% of the total variance, thereby establishing an investor sentiment measurement model. Further, to establish a correlation between trading volume and investor sentiment, we conducted logistic regression analysis based on the sentiment measurement model. The tested model achieved an accuracy rate of 87.3%. Additionally, using a Python web crawler, we collected and standardized the "greed and fear index" as an emotional indicator, matched it with daily data, and incorporated it into the logistic regression model. This enhanced model exhibited an accuracy rate of 82.5%. Our findings reveal that the greed and fear index has a more significant impact on emotional index changes than trading volume and turnover rate in logistic regression. This work not only contributes to a deeper understanding of investor sentiment in the non-ferrous metal stock market but also provides insights into potential market behaviors, guiding investors and policymakers in making informed decisions.