Application of Deep Learning in Stock Market Valuation Index Forecasting

Abstract

Deep learning is the core technology of artificial intelligence, which has higher accuracy than traditional algorithms. The characteristics of high-risk and high-yield in stock market make investors hope to make predictions on it through scientific methods, so as to reduce investment risks. Long short-term memory (LSTM) model in deep learning can effectively describe the long memory of data and is suitable for predicting financial time series. Therefore, this paper uses LSTM model in deep learning to learn and forecast the stock market valuation indicator, price-earnings ratio (P/E ratio). Then the prediction bias is measured by forecast trend accuracy (FTA), average forecast deviation rate (AFDR), and root mean square error (RMSE). Empirical results show that LSTM model has a good predictive effect on P/E ratio sequence, indicating that there is practical research value for applying deep learning network algorithm to the field of stock market forecasting. At the same time, this paper also provides a reference for stock market investors.