Using Deep Learning To Predict Stock Price Based On A+H Shares Empirical Analysis

With the frequent financial transactions between the mainland and Hong Kong,more people have invested in the analysis of the H-share market.However,if we look at the existing research results,it is not comprehensive.A+H shares refer to the listing on the Shanghai Stock Exchange or the Shenzhen Stock Exchange not only in the A shares but also on the corresponding exchanges in Hong Kong as H shares.That is,domestic companies have launched stocks in mainland China and Hong Kong.The price of A-share H-shares is not only affected by the external economic and political background.Man-made events are also closely related to the overall psychological quality and knowledge level of investors.The A-share H-share academic issue remains to be further analyzed by scholars in all aspects in order to unveil its mystery.In today's globalization,more and more mainland companies choose to launch H-shares in Hong Kong.Therefore,the analysis of H-shares of A-shares can indeed help solve practical problems.Deep learning is based on machine learning.It enhances the learning ability through multiple hidden layers so that the features are better expressed and the accuracy of model prediction is greatly improved.In 2016,Google AlphaGo artificial intelligence defeated,one of the world's best players in the Go tournament,in the Go tournament.The advantage was obvious and caused intense attention.Deep learning triggered an upsurge in various fields,including financial data prediction.Apply deep learning to A-share H-share research.This paper selects 8 bank shares in 40 constituent stocks of the Heng Sheng China Enterprise Index for analysis.According to the order of stock numbers,it is followed by CCB(601939.SH,0939.HK),CITIC Bank(601998.SH,0998.HK),and Agricultural Bank(601288.SH,1288.HK),ICBC(601398.SH,1398.HK),Minsheng Bank(600016.SH,1988.HK),Bank of Communications(601328.SH,3328.HK),China Merchants Bank(600036.SH),3968.HK),Bank of China(601988.SH,3988.HK).First describes the data,and describes the data at the same time,describes the development history and current status of A-share H-shares,and then selects the model after understanding the data.The model is a long-term and short-term memory neural network LSTM developed on the basis of a recurrent neural network.The LSTM predicts the prices of A-shares and H-shares,describes the prediction results in detail,and revises the results to further revise the model.For the improved model to re-predict the original data,compare the changes before and after the model improvement,and compare the A share H shares with the difference and find in which market the prediction model better.