Using Deep Learning Neural Networks and Stacking Ensemble Learning to Predict CSI 300 Index

In the digital age, more and more households begin to invest in the stock market. Deep learning has developed rapidly in recent years. Deep learning-based methods have also been applied to the securities investment market. Stacking is an ensemble learning algorithm, which synthesizes predictions from multiple machine learning algorithms to produce more accurate predictions. Temporal Convolutional Network (TCN) is a neural network model that uses causal convolution and dilated convolution to process time series. Long Stort-Term Memory (LSTM) is a kind of Recurrent Neural Network (RNN) that can carry information across multiple time steps and performs well in time series prediction. Gated Recurrent Unit (GRU) is another kind of RNN, which has a different structure from LSTM, but can also achieve good results in time series prediction. In addition, Convolutional Neural Network (CNN) is good at extracting the characteristics of data from multidimensional sequences. In order to better solve the stock prediction problem, this paper integrates TCN, CNN-LSTM and GRU through stacking to get an ensemble learning model. This paper selects the data of CSI300 Index and uses the data of the first 60 trading days as characteristics to predict the opening price of this trading day. Experiments show that the proposed algorithm and model can achieve higher accuracy than TCN, CNN-LSTM and GRU in stock prediction.