Design And Implementation Of Stock Closing Price Prediction System Based On PCA-CNN-IGRU

With the development of the stock market,stock investment has become one of the most popular investment methods in society today.Accurate stock price prediction is of great significance for promoting market supervision and assisting investors in avoiding risks.Stock price prediction is a typical nonlinear time series predicting problem,and stock price changes are affected by many factors.However,too many influencing factors will lead to redundant input of the prediction model and increase the calculation amount of the model.With the introduction and development of the neural network,it is gradually applied to stock price prediction and has achieved good results.Therefore,this paper proposes a stock closing price prediction model based on principal component analysis(PCA),convolutional neural networks(CNN)and improved gated recurrent unit(IGRU),PCA-CNN-IGRU.PCA reduces the model input redundancy and the calculation amount of the neural network by reducing the dimension of input data,thus improving the model calculation efficiency.CNN can extract important features of data.IGRU is an improved model of gated recurrent unit(GRU),which further alleviates the occurrence of gradient explosion and gradient disappearance by introducing the data conversion module(DCM)after calculating the reset gate.Therefore,the IGRU model can better learn the long-term dependency between data and improve the prediction accuracy of the model.This paper takes the Shanghai Composite Index as the research object.The experimental data include the historical trading data of the Shanghai Composite Index and the closing prices of the Nasdaq Index,the Dow Jones Industrial Average,the Hang Seng Index and the Shenzhen Component Index from July 15,1991,to December 30,2022.In this paper,support vector regression(SVR),long short-term memory(LSTM),GRU,IGRU,PCA-LSTM,PCA-GRU,PCA-IGRU,PCA-CNN-LSTM,PCA-CNN-GRU are used as baseline models for the comparative experiment.The experimental results show that the PCA-CNN-IGRU model has higher prediction accuracy than the other models.This paper completes the design and implementation of the stock closing price prediction system based on the PCA-CNN-IGRU model.The system realizes the functions of stock data query and stock closing price prediction,which can provide an effective reference for users.