S&P 500 Index Value Forecasting Using Decision Fusion Regression Model

Stock information like the S&P500 index is extensive, intricate, nonlinear, and filled with disturbances. Hence, finding appropriate methods to forecast its value has emerged as a complex task. This study uses a regression approach to explore the potential of a BiLSTM-GRU decision fusion model. The individual models exhibit promising outcomes during training, testing, and validation. A fusion option enhances the forecast accuracy for new extrapolation scenarios. The new model shows good predictive performance for short-term forecasting and improving outcomes compared with studies utilized for benchmarking. This focus offers an avenue for academics to delve into a novel exploration of data prediction. Simultaneously, professionals can leverage this research to refine criteria supporting investment choices within the stock market.