Deep learning of predicting closing price through historical adjustment closing price

Abstract

In recent years, deep learning has shown its power in many fields such as computer vision and natural language processing. However, there is not much research on stock prediction using deep learning, so we explored and proposed a deep learning model based on LSTM to predict stock closing prices. The model can predict the closing price of the stock on the following day based on the previous day’s total volume, the opening price, and the adjustment closing price that can reflect the peripheral information of the stock. To improve the accuracy of the prediction, we invoke the attention module on the model. Experiments show that our model performs well on two datasets, the CSI 300 and the Hang Seng Index, with very little error between the predicted closing price and the actual closing price.