An Interpretable Framework for Stock Market Forecasting Using Long Short-Term Memory (LSTM) Networks with SHAP - Explainable AI (XAI) Method

Predicting stock market movements is a challenging and ever-changing process because it is by a variety of unpredictable elements. In this study, we explore the use of deep learning, specifically Long Short-Term Memory (LSTM) networks, to estimate the upcoming stock prices of HDFC Bank. The training process of the model utilizes past stock closing prices to learn patterns and make future predictions from 2020 to 2025, sourced from Yahoo Finance. The time series data is preprocessed using normalization techniques, and sequences of past observations are structured to train the LSTM architecture effectively. To make the model's predictions easier to understand and explain, Explainable AI methods, particularly SHAP (SHapley Additive exPlanations), are employed. The proposed model demonstrates strong predictive performance, achieving a Root Mean Squared Error (RMSE) of 0.0246, indicating that the average difference between the predicted and actual values is quite low. Additionally, the R² (coefficient of determination) score was 0.9489, suggesting that approximately 94.89% of the variability in the actual values is explained by the model's predictions. These results highlight the model's capability to accurately capture trends in stock price movement while also providing valuable insights into feature contributions. Unlike traditional statistical models, LSTM networks can retain long-term dependencies, enabling them to detect time-based patterns effectively. The integration of explainability not only adds transparency to the predictions but also enhances the trustworthiness of the model in real-world applications. Furthermore, this approach can be adapted to other financial instruments or extended with external indicators for broader market analysis. The combination of LSTM and SHAP enables a robust, interpretable framework for stock market forecasting that supports data-driven investment strategies.