| The Efficient Market Hypothesis has laid an important foundation for the development of traditional financial theory,which describes the operation state of the financial market under a rational framework.However,all kinds of anomalies that followed showed that the reality was not as described in the theory,but the changes of the stock market did not meet the requirements of random walk.Many factors at different levels were affecting the formation of the stock price,and it would be profitable to predict the changes of the stock market based on these factors and adopt corresponding investment strategies.The development of behavioral finance has strengthened this research trend,and irrational factors have attracted more and more attention from relevant scholars.More and more research shows that the psychological cognitive bias and subjective emotions of market participants can have a significant impact on the future trend of securities prices,and the use of these subjective emotional information can better predict market indicators such as stock prices.Among various market anomalies,momentum and reversal are paid more attention to.Behavioral finance school has made some explanations for them from the perspective of investor sentiment,but they have not been unanimously recognized by the academic community.Many scholars,starting from the theoretical framework of traditional finance,believe that a few optimized asset pricing models can also explain the emergence and extent of momentum and reversal effects.Although there are many literatures about momentum effect at present,few studies have quantified its degree and further explored the predictive effect of this indicator on the stock market trend.After sorting out and summarizing the above aspects,this paper believes that investor sentiment and market momentum have an important impact on stock market changes,so it is added to the prediction model as an input variable to analyze whether it can improve the prediction effect.Specifically,this paper collected the relevant data of more than 2000 stocks in the Shanghai A-share market on all trading days from January 1,2015 to December 31,2021,and used certain indicator construction methods to build investor sentiment indicators and momentum and reversal indicators that reflect market momentum.The prediction effects of these two indicators were studied respectively in terms of stock index prediction and market volatility prediction,In the aspect of stock index prediction,the basic model is the model that only uses price volume data to predict the daily closing index,and in the aspect of market volatility prediction,the basic model is the model that only uses the volatility’s own time series data to predict.On the basis of such basic models,it is explored whether the prediction accuracy and other evaluation indicators have significantly improved after adding sentiment and momentum indicators.The prediction model used in this paper is LSTM long and short term memory artificial neural network model which is considered to be good at processing long time series data.Through comparative analysis,this paper mainly draws the following three conclusions: First,compared with the basic models using traditional prediction methods,the prediction models incorporating investor sentiment information can achieve better prediction results in terms of both stock index prediction and market volatility prediction,which to some extent also verifies the conclusion that investor sentiment has a significant impact on stock market trends in previous studies;Second,the performance of the model with momentum factor and reversal factor is also better than the basic models in these two aspects.Although the improvement effect of volatility prediction is limited,the overall view can also explain to a certain extent that the information on market momentum is also helpful to predict the stock market trend;Third,by adding the combination of emotion and momentum at the same time,the prediction accuracy has improved on the basis of the first two situations,which provides a new perspective for the research in the field of stock market prediction.The main contribution of this paper is that it has made attempts in the measurement and quantification of momentum factors that few people have studied at present,and further analyzed the impact of these two factors on the prediction of stock index and volatility in combination with investor sentiment,which has certain theoretical significance for the exploration of effective prediction factors in related fields. |
