Research Article
Prediction of Shanghai Stock Market Based on CNN-LSTM Model with GA Optimized Attention
@INPROCEEDINGS{10.4108/eai.19-5-2023.2334242, author={Hanyu Hu}, title={Prediction of Shanghai Stock Market Based on CNN-LSTM Model with GA Optimized Attention}, proceedings={Proceedings of the 2nd International Conference on Bigdata Blockchain and Economy Management, ICBBEM 2023, May 19--21, 2023, Hangzhou, China}, publisher={EAI}, proceedings_a={ICBBEM}, year={2023}, month={7}, keywords={stock predictions cnn-lstm attention mechanism genetic algorithm}, doi={10.4108/eai.19-5-2023.2334242} }- Hanyu Hu
Year: 2023
Prediction of Shanghai Stock Market Based on CNN-LSTM Model with GA Optimized Attention
ICBBEM
EAI
DOI: 10.4108/eai.19-5-2023.2334242
Abstract
Investment in stocks is a means of maximizing the benefits of assets through the purchase and sale of stocks. For equity investment, accurate forecasting of stock prices, and ups and downs are essential for asset return management. Currently, there are not only classical Markov chain models, but also many models with deep learning for predicting time-series outcomes. By studying the current relevant work on stock forecasting, it is recognized that the existing forecasting models have a low number of feature extractions from stock data and insufficient training data, resulting in a lack of forecasting accuracy of the models. For improvement of the prediction accuracy, this paper proposes a CNN-LSTM model with GA-optimized Attention, which is a combined prediction model based on the traditional LSTM model with a convolutional module to extract stock features, an Attention mechanism to improve accuracy, and a GA algorithm to find the optimal fully connected layer weights. In addition, this paper uses the Qlib platform to calculate the stock Alpha158 factor in data processing, and the RFE and PCA methods for feature reduction and extraction to obtain the processed variables.
