Stock Price Prediction On New Energy Vehicle Industry Chain Based On GSRNN Model

Deep learning models generally have higher prediction accuracy than traditional time series models,so they are widely used in the financial field.In recent years,a large number of investors have emerged in the new energy vehicle stock market,so this paper uses a deep learning composite model to predict the relevant data of the new energy vehicle industry chain.In order to help investors accurately grasp the development trend of the new energy vehicle upstream,midstream and downstream industries,this paper compiled the new energy vehicle upstream,midstream and downstream industry index from July 1,2017 to July 1,2021,and compared The index and the closing prices of two groups of six stocks with higher average daily turnover and market value in the new energy vehicle industry chain are predicted.Considering the correlation of these data,the recurrent neural network based on the weighted directed graph structure is selected.(Graph structured recurrent neural network,GSRNN)model for prediction.In the weighted directed graph structure,the nodes are indices or stocks,the association edges are the association relationships between nodes,and the association edge weights are the association sizes between the nodes.In the analysis method of correlation relationship,gray correlation analysis has lower requirements on sample size and distribution law,and the correlation coefficient calculated by Pearson correlation coefficient method is not affected by the location and scale of variables.Therefore,this paper uses Pearson correlation coefficient method and gray correlation method.Analysis Two methods are used to perform association analysis on nodes.If the correlation coefficient and gray correlation degree between nodes are greater than 0.6,it is considered that there is a bidirectional correlation edge between nodes,and the weight of the correlation edge is the standardized correlation coefficient and gray correlation degree.This paper takes the new energy vehicle upstream,mid-stream and downstream industry indices and the closing price time series of six stocks from July 1,2017 to July 1,2021 as the research objects,and divides the data into training set and test set.Use the training set to train the GSRNN model and use the test set to make predictions.The prediction accuracy of the model is judged by comparing the size of the model evaluation indicators of the GSRNN model,ARIMA model and LSTM model on the test set.The results show that the GSRNN model has obvious advantages in prediction.In order to illustrate the scalability of the GSRNN model,the paper finally predicts the representative stock price indices in the domestic financial market,and the prediction results also prove the effectiveness of the GSRNN model.