Stock prediction based on random forest and LSTM neural network

Abstract

The data in the stock market are intricate. Principal Component Analysis (PCA) based on LSTM neural network can remove noise and improve the accuracy of stock prediction. A stock prediction model based on random forest and LSTM neural network is proposed to further improve the performance of stock prediction. Based on the data of Shanghai Composite Index from 2013 to 2017, this model and PCA + LSTM neural network model are simulated and compared. The experimental results show that this model is more suitable for stock prediction than PCA + LSTM model. In addition, the returns of trading strategies based on the above two models are higher than the benchmark buy-and-hold strategy, and the trading strategies based on the proposed model perform best.