Stock market trend forecasting using gated recurrent unit deep learning model

Stock market forecasting is extremely risky for stockholders and economic analysts. This study aims to explore the correlation between the historical stock data and future stock prices of four prominent companies based on Open, High, Low, Close, and Volume (OHLCV) indicators. To capture temporal dependencies in time-series data, Gated Recurrent Unit (GRU)-based deep learning model is employed. The suggested framework combines the classic stock market analysis with deep learning methods for better forecasting performance to assist investment decision making. The experimental results have shown that the model can accurately forecast the stock’s price and direction of the trend, with the MAPE ranging from 2.1% to 3.5% and the direction accuracy ranging from 73% to 85% for all stocks. The results indicate that GRU-based forecasting can be beneficial to investors and analysts. Future studies should, however, be directed towards comparative benchmarking with other forecasting models, as well as towards adding more market indicators to make the model more robust and to enhance the generalizability.