An Empirical Study On Comprehensive Stock Selection Based On Principal Component Analysis And BP Neural Network

With the rapid development of financial industry and computer technology,quantitative stock selection research to obtain excess or stable investment income has gradually emerged in our country.Therefore,the stock market is a complex nonlinear system.It is very difficult to construct the model as the tool of stock price forecasting and investment guidance.And the traditional quantization model is partial.Therefore,this paper will be based on the principal component analysis of listed companies to assess the financial status of the study,based on principal component analysis and BP neural network combination of stock price forecasting and traditional stock selection analysis method to build a diversified quantitative stock selection model.Because neural network has a unique advantage in dealing with nonlinear problems in recent years,neural network through the study of stock history data,the law of stock price changes in the weight of neurons,through the use of trained networks to predict the future share price.The main work of this paper is as follows:Firstly,the main financial indicators of listed companies in Shenzhen A-share market are analyzed by using principal component analysis(PCA),and the comprehensive score of evaluation index is obtained.Based on the comprehensive score,the comprehensive and comprehensive understanding of listed companies Financial status,empirical study shows that the effectiveness of the model evaluation method.Secondly,this paper uses the financial index data and the transaction index data as the input variables of the stock price forecast research.The empirical study finds that the stock price forecasting based on the method of principal component analysis and BP neural network can effectively and accurately evaluate the stock next year In the same quarter of the ups and downs,we used three BP neural network algorithm to carry out experiments,which based on the Bayesian regularization algorithm BP neural network to build the model to predict the highest accuracy and model training process can effectively avoid the over-fitting phenomenon,The predicted value and the actual value of the error is small.Finally,the comprehensive stock selection scheme is as follows: firstly,the top25% of the stocks recommended by the Institute of Financial Position Evaluation of listed companies are selected based on the principal component analysis method,and then the stock price forecast based on principal component analysis and BP neural network is selected Institute of the top 25% of the stock recommended,and finally combined with the traditional stock selection analysis method of comprehensive analysis of listed companies to determine the financial status of the study and stock price forecast both recommended by the top 25% of the stock in the small cap stocks as the next year The best investment in the year.