Multiple-Branch Convolutional Neural Network for SSEC Daily Return Prediction

Shengxi Jiang

2025

Abstract

This study proposes a Multiple-branch Convolutional Neural Network (MBCNN) model to predict the daily return direction of the Shanghai Securities Composite Index (SSEC). Due to the complexity and volatility of financial markets, traditional machine learning methods such as Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), and Decision Tree (DT) often face limitations in capturing intricate patterns. To address this, the model leverages multiple feature branches, incorporating historical price data, global market indicators, and other financial metrics. The effectiveness of MBCNN is evaluated against classical machine learning approaches, with results demonstrating superior performance in both accuracy and F-measure metrics. Additionally, the study explores the impact of Principal Component Analysis (PCA) on model performance, revealing that PCA does not enhance prediction accuracy. Experimental results confirm that MBCNN outperforms traditional models, offering improved classification capabilities and robustness. These findings provide valuable insights and a foundation for future research on stock market trend prediction.

Paper Citation

in Harvard Style

Jiang S. (2025). Multiple-Branch Convolutional Neural Network for SSEC Daily Return Prediction. In Proceedings of the 2nd International Conference on Data Science and Engineering - Volume 1: ICDSE; ISBN 978-989-758-765-8, SciTePress, pages 265-270. DOI: 10.5220/0013686200004670

in Bibtex Style

@conference{icdse25,
author={Shengxi Jiang},
title={Multiple-Branch Convolutional Neural Network for SSEC Daily Return Prediction},
booktitle={Proceedings of the 2nd International Conference on Data Science and Engineering - Volume 1: ICDSE},
year={2025},
pages={265-270},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013686200004670},
isbn={978-989-758-765-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Science and Engineering - Volume 1: ICDSE
TI - Multiple-Branch Convolutional Neural Network for SSEC Daily Return Prediction
SN - 978-989-758-765-8
AU - Jiang S.
PY - 2025
SP - 265
EP - 270
DO - 10.5220/0013686200004670
PB - SciTePress