Research On Stock Price Momentum Investment Strategy From The Perspective Of HS-IPS Mode

As one of the most classic models in the field of behavioral finance,HS model(Hong-Stein model)has always been the research focus in this field.In recent years,a large number of studies have discussed the decision-making behavior of stock investors based on HS model,but the research from the perspective of information dissemination speed is not enough.At the same time,the significant progress of deep learning methods has greatly promoted the popularization of artificial intelligence applications.The new data dimension brought by Internet user behavior also provides a new perspective for stock price prediction.In the stock market with many influencing factors and great uncertainty,artificial intelligence based on deep learning is more and more used in quantitative investment.As an improved recurrent neural network,LSTM(long short term memory)is good at exploring the relationship between nonlinear data.It can find out the influence between time series through the learning of historical relevant data,and deeply mine the old rules of stock price through the selective memory function of machine learning,so as to predict the information propagation speed(IPS)With the addition of new variables such as investor concern and momentum,LSTM can refer to more comprehensive investor behavior and decision-making information,which provides a new opportunity to improve the prediction ability of recurrent neural network.Combined with the data of different dimensions such as CSI 300 index,CSI 500 index,26 industries,20 stocks,Baidu search index and Oriental Wealth attention,this paper first constructs the information dissemination speed model,introduces the speed variable to improve the HS model,and constructs the HS-IPS model,Granger test and correlation analysis are used to test the correlation and causality between various variables and stock closing price;Secondly,LSTM neural network is selected to verify the above research results.Finally,the adaptability of LSTM neural network is improved according to the characteristics of different data,and the quantitative investment strategy is compared.