Deep Learning for Stock Market Prediction: A Comparative Study on Nepal’s Commercial Banking Sector
DOI:
https://doi.org/10.26821/IJSHRE.13.10.2025.131009Keywords:
Deep Learning, NEPSE, LSTM, Transformer, TimesNetAbstract
The prediction of stock market trends remains complicated in emerging markets like Nepal because of frequent market volatility and numerous economic influences. This research explores how machine learning and deep learning algorithms function in predicting stock market values for commercial banks that trade on the Nepal Stock Exchange (NEPSE). Daily historical stock data were collected from the time frame 2019-2024, along with external financial indicators such as gold prices, exchange rates, fuel prices, inflation rates, and interest rates, which were also collected from the same period and were preprocessed through normalization, missing value imputation, and interpolation for non-daily indicators. The research tests Long Short-term Memory (LSTM) along with Transformer and TimesNet using Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and Mean Absolute Error (MAE) evaluation methods.
The TimesNet model achieves superior performance than both LSTM and Transformer by delivering a 30–50% reduction in RMSE, MSE, and MAE across various commercial banks. These findings highlight the advantages of foundation models in financial forecasting, offering more accurate and stable predictions compared to traditional deep learning methods. The research findings provide essential knowledge to investors, policymakers, and financial analysts who need it for informed decision-making in the Nepalese financial sector. Future studies could explore hybrid models and additional. macroeconomic variables to enhance predictive performance further.
References
. International Monetary Fund, “Nepal: Staff Report for the 2023 Article IV Consultation,” 2023.
. World Bank, “Nepal Financial Sector Assessment,” 2023.
. H. KC, “Performance analysis and prediction of Nepal stock market (Nepse) for investment decision using machine learning techniques,” International Journal of Computer Science Engineering (IJCSE), vol. 7, no. 1, pp. 15–27, 2018.
. N. R. Pokhrel, K. R. Dahal, R. Rimal, H. N. Bhandari, R. K. C. Khatri, B. Rimal, and W. E. Hahn, “Predicting NEPSE index price using deep learning models,” Machine Learning with Applications, vol. 9, p. 100385, 2022.
. M. Gurung, P. Neupane, and S. Shrestha, “Stock market prediction in Nepali stock market using machine learning model,” KEC Journal of Science and Engineering, vol. 8, no. 1, pp. 158–168, 2024.
. H. Wu, T. Hu, Y. Liu, H. Zhou, J. Wang, and M. Long, “TimesNet: Temporal 2D-variation modeling for general time series analysis,” in Proc. Int. Conf. Learn. Representations (ICLR), 2023.
. M. Mallikarjuna and R. P. Rao, “Evaluation of forecasting methods from selected stock market returns,” Financial Innovation, vol. 5, no. 1, p. 40, 2019.
. G. S. Atsalakis and K. P. Valavanis, “Surveying stock market forecasting techniques–Part II: Soft computing methods,” Expert Systems with Applications, vol. 36, no. 3, pp. 5932–5941, 2009.
. D. Kumar, P. K. Sarangi, and R. Verma, “A systematic review of stock market prediction using machine learning and statistical techniques,” Materials Today: Proceedings, vol. 49, pp. 3187–3191, 2022.
