Investigating the Relationship between Business Sector Returns in the Japanese Stock Market and its Impact on Forecasting Returns by CNN-LSTM.

Abstract

This study investigates the relationship between business sector returns and stock price forecasting in the Japanese stock market. First, we employ the vector autoregression with exogenous variables (VARX) and the Granger causality test to analyze inter-sectoral linkages and their impact on sectoral returns. We propose a hybrid machine learning approach, the CNN-LSTM model, which incorporates the abovementioned assumptions for improved prediction performance. Our findings indicate that considering inter-sectoral relationships enhances the accuracy of stock price forecasting.