GraphCNNpred: A stock market indices prediction using a Graph based deep learning system
Yuhui Jin
TL;DR
This work addresses the challenge of stock market index prediction by leveraging a graph-based extension of CNNpred, termed GraphCNNpred, which integrates graph neural networks with CNN-based feature extraction to utilize diversified data. By constructing a feature-correlation graph with edges for correlations above a threshold and implementing two main variants (GAT-CNNpred and GCN-CNNpred) along with graph pooling, the authors achieve higher predictive performance across the S&P 500, NASDAQ, DJI, NYSE, and RUSSELL indices. They demonstrate improvements of about $4\%$ to $15\%$ in $F$-measure over strong baselines and show that trading simulations yield a Sharpe ratio exceeding $3$, indicating practical trading viability. The approach combines daily feature aggregation via graph layers with durational pattern extraction through CNN components, providing a robust framework that effectively exploits inter-feature relationships and temporal dynamics for multi-index prediction.
Abstract
The application of deep learning techniques for predicting stock market prices is a prominent and widely researched topic in the field of data science. To effectively predict market trends, it is essential to utilize a diversified dataset. In this paper, we give a graph neural network based convolutional neural network (CNN) model, that can be applied on diverse source of data, in the attempt to extract features to predict the trends of indices of \text{S}\&\text{P} 500, NASDAQ, DJI, NYSE, and RUSSEL. The experiments show that the associated models improve the performance of prediction in all indices over the baseline algorithms by about $4\% \text{ to } 15\%$, in terms of F-measure. A trading simulation is generated from predictions and gained a Sharpe ratio of over 3.