Research On Stock Price Prediction Model Based On CCA-GA-BPNN Comprehensive Technology

In recent years,with the vigorous development of the market economy and the continuous improvement of the financial system reform,the scale of China's stock market has continued to grow and developed,and has developed into an important part of the market economy.More and more Chinese people are beginning to pay attention to the stock market and participate directly or indirectly in stock market investment.Discussions on the stock market have also become part of people's daily lives.The scientific modeling of stock data to realize stock price prediction has become a major reality.And theoretical work.Due to the intricacies of stock market movements,traditional statistical learning methods are all based on linear models,and they cannot learn the internal laws that affect stock prices well.The neural network model can learn the nonlinear mapping relationship from the data,which is more practical and effective than the traditional method,and has great research significance.In order to fully consider the factors that affect the stock market,this article selects 16 indicators for stock price forecast,including 6 market indicators including opening price,closing price,highest price,lowest price,turnover,and trading volume,including market value and turnover rate?PB,PC,PE,PS total 6 market value indicators,including 21-day moving average MA21,smoothed moving average MACD,20-day closing price standard deviation 20 SD,index moving average EMA total 4 Technical indicators.Due to the problem of information redundancy between these features,the author uses typical correlation analysis to reduce dimensionality,and selects several representative typical variables as input to the neural network,which reduces the number of nodes in the network input layer and shortens the network training time.Since the neural network is easy to converge to the local optimal value,in order to improve the prediction accuracy of the neural network,the genetic algorithm is used to optimize the weights and thresholds of the neural network structure,so that the network prediction is easier to approach the global optimal solution,and at the same time improve the model generalization ability.This article takes the daily data of Ping An Bank and Founder Technology as the research object,which elaborate on the construction of the optimal model based on Ping AnBank's stock data,with mean square error(MSE)and relative error(MAPE))Is the evaluation index of stock price prediction effect,and the model is used to predict the closing price of the next day.The results of matlab simulation experiments prove that the neural network model(CCA-GA-BPNN)based on typical correlation analysis and dimension optimization combined with genetic algorithm optimization is more predictable than BP model,CCA-BP model,GA-BP model,PCA-GA-BPNN model The effect is good,and the network structure is relatively simple,which can reduce risks and increase returns.