A Stock Price Prediction Approach Based on Time Series Decomposition and Multi-Scale CNN using OHLCT Images

TL;DR

This paper introduces a novel deep learning model that combines time series decomposition with multi-scale CNNs on OHLCT images to improve stock price prediction accuracy in the China A-share market, demonstrating significant profit gains.

Contribution

The paper presents a new multi-scale fusion CNN model that integrates sequential and image features, reducing overfitting and enhancing prediction accuracy for stock movements.

Findings

Achieved 61.15% positive predictive value for 5-day stock trend prediction.

Attained 63.37% negative predictive value for 5-day stock trend prediction.

Generated a total profit of 165.09% in experiments.

Abstract

Recently, deep learning in stock prediction has become an important branch. Image-based methods show potential by capturing complex visual patterns and spatial correlations, offering advantages in interpretability over time series models. However, image-based approaches are more prone to overfitting, hindering robust predictive performance. To improve accuracy, this paper proposes a novel method, named Sequence-based Multi-scale Fusion Regression Convolutional Neural Network (SMSFR-CNN), for predicting stock price movements in the China A-share market. By utilizing CNN to learn sequential features and combining them with image features, we improve the accuracy of stock trend prediction on the A-share market stock dataset. This approach reduces the search space for image features, stabilizes, and accelerates the training process. Extensive comparative experiments on 4,454 A-share stocks show that the model achieves a 61.15% positive predictive value and a 63.37% negative predictive value for the next 5 days, resulting in a total profit of 165.09%.