Research On The Relationship Between Investor Sentiment And Stock Market Index Volatility Based On Text Mining

The volatility of stock market index is what securities investors focus on,and its impact factors cover a wide range of aspects,with directly related impact factors and some potential factors that are difficult to detect.In recent years,with the strengthening of computer computing power and the emergence of many artificial intelligence algorithms,more and more scholars and investors have combined artificial intelligence algorithms with the laws of the stock market for stock price volatility prediction research,and applied them to quantitative investment and stock market regulation and other aspects with significant results.This paper explores the interrelationship between emotional tendencies in speech messages and stock index changes to construct a comment-based stock index prediction model from the perspective of shareholder speech.In order to apply irrational decision factor investor sentiment to stock market stock index volatility prediction,this paper integrates deep learning algorithm to construct a more accurate stock index volatility prediction model,and proposes an emotion propensity refining method based on natural language processing and convolutional neural network,to construct investor sentiment metrics by combining emotion indicators with stock market indicators,and to predict future stock index volatility using stockholder sentiment data combined with an improved LSTM model(ELSTM).The study reached the following conclusions.(1)The indicator of investor sentiment based on stock BBS commentary can reflect the change of stock index,and the OLS regression test proves that this paper constructs a valid expression of compound sentiment of investor sentiment.(2)On the premise of excluding the influence of policies,natural disasters and the financial environment,a comprehensive analysis of the changes in investor sentiment and the stock price index shows that there is no positive or negative correlation between the changes in investor sentiment and stock price in the stage of stock price rise;while in the stage of stock price fall,the composite investor sentiment index is predictable for the changes in stock price,and the index has a positive correlation with sentiment.(3)The ELSTM model,which incorporates sentiment indicators,has significantly improved its prediction accuracy in stock market forecasting,and can make more accurate predictions onthe trend of CSI 300,which is better than other models,with a prediction accuracy of 85.4% on the trend of CSI index and 64.3% on the overall prediction of different stock sectors.The model's viability is demonstrated by the substantial improvement in accuracy compared to the ordinary time series model and the LSTM model.It has a predictive effect on the prediction of the trend of relatively inactive industry concept stocks,providing investors with new solutions for investment and stock market analysis,and providing a new way for stock market index prediction.