Aiming at the problem that deep features of stock data are difficult to extract and the prediction accuracy is not high, an improved LSTM model CGLA is constructed. Firstly, the RNN-Attention model, LSTM-Attention model and GRU-Attention model are constructed by using attention mechanism. GRU-Attention model with the best performance is selected by comparison. The deep features of stock time series data are extracted by CNN and sent to GRU-Attention model. Then LSTM is used to improve the network structure of the above training model, based on this, a hybrid CGLA model (CNN-GRU-LSTM-Attention) is constructed to predict the stock price of CSI300. After experimental verification, the MSE of CGLA model is reduced by two orders of magnitude compared with the comparison model, the R2_score is significantly improved, and the running time of CGLA model is greatly shortened compared with the comparison model. This paper also integrated factor correlation analysis, in a number of stock indicators in a comprehensive analysis of the closing price of the relevant stock indicators, combined with CGLA model to predict. The experimental results show that the combination of deep learning model and stock index influence factors can make the experiment obtain more accurate and more real stock trend prediction results.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.