Multi-factor Quantitative Stock Selection Model Based on LSTM-TabNet

With the growing maturity and rapid development of artificial intelligence technology, more and more scholars and institutions at home and abroad are applying it to the field of financial investment, especially in quantitative stock selection. In order to further improve the accuracy of stock price prediction and stock selection, this article proposes a stock selection model based on the combination of LSTM and TabNet. For A-share stock data, a LSTM stock selection prediction model and a TabNet stock selection prediction model are first established. Then, factors positively related to stock returns are obtained through financial knowledge, and finally, the factors are combined with the model that fuses LSTM and TabNet algorithms for stock selection prediction. By utilizing the mature processing capability of LSTM for handling long sequence data with changes and the excellent performance of TabNet in handling table-type data, combining the two can enable the quantitative stock selection model to achieve a high Sharpe ratio and excess returns, providing reference for financial investors to choose high-quality stocks.