Enhancing Stock Price Prediction Through Fusion of Deep Learning Models and Technical Indicators

Accurate forecasting of stock price movements is important for traders and investors to handle the unpredictable nature of the financial market. In this study, deep learning models are integrated with Technical Indicators (TI) for stock price prediction using Tata Consultancy Services (TCS) and Infosys Ltd (INFY) over ten years, from 31/07/2014 to 31/07 /2024. We gathered daily stock data on opening, closing, lowest, highest prices and trading volume to monitor the market's daily performance. The opening price was selected as the target variable to understand price movements and overall market behavior. Along with five TI, two deep learning methods, long short-term memory (LSTM) and gated recurrent unit (GRU), were employed. The technical indicators included simple moving average (SMA), exponential moving average (EMA), moving average convergence divergence (MACD), relative strength index (RSI), and bollinger bands (BB). To measure the performance of our model, we used R-squared (R2), root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). The results show that the proposed LSTM - GRU model, using OHLCV and TI, significantly improves prediction accuracy compared to other models, achieving R2 values of 0.981 for the TCS dataset and 0.976 for the Infosys dataset. Combining deep learning with technical analysis proves more effective for predicting stock prices than using either method alone. This approach helps investors and traders by providing a more accurate tool for understanding the market dynamics and making profitable decisions.