Title: Volatility modelling and forecasting in stock markets: a machine learning approach

Authors: Soumen Ghosh; Kuntal Mukherjee; Biswajit Jana; Syed Saif Ahmed; Mohammad Aasif; Sayel Munsi

Addresses: Haldia Institute of Technology, Department of Computer Science and Engineering (Data Science), Haldia, 721657, West Bengal, India ' Haldia Institute of Technology, Department of Computer Science and Engineering (Data Science), Haldia, 721657, West Bengal, India ' Department of Computer Science and Engineering, Bennett University, Greater Noida, 201310, Uttar Pradesh, India ' Haldia Institute of Technology, Department of Computer Science and Engineering (Data Science), Haldia, 721657, Pin Code-721657, West Bengal, India ' Haldia Institute of Technology, Department of Computer Science and Engineering (Data Science), Haldia, 721657, Pin Code-721657, West Bengal, India ' Haldia Institute of Technology, Department of Computer Science and Engineering (Data Science), Haldia, 721657, Pin Code-721657, West Bengal, India

Abstract: This research explores the application of various models for stock price prediction, including ARIMA, LSTM, SARIMAX, and a hybrid SARIMAX-LSTM, highlighting their importance in the post-pandemic financial landscape. The study emphasises the limitations of traditional methods and the necessity of time-series analysis for understanding stock price patterns. It focuses on the impact of COVID-19 on financial markets and assesses the reliability of these models in unpredictable conditions. The methodology involves data selection, pre-processing, model parameter tuning, and performance evaluation. The research establishes a framework for the implementation of these models, underscoring the need for parameter optimisation to enhance accuracy. Ultimately, the study shows that LSTM performs better than the other models and offers valuable insights into using advanced forecasting techniques for improved investment strategies in the evolving stock market.

Keywords: LSTM; long-short term memory; ARIMA; autoregressive integrated moving average; SARIMAX autoregressive (AR); stock market.

DOI: 10.1504/IJDATS.2025.150912

International Journal of Data Analysis Techniques and Strategies, 2025 Vol.17 No.4, pp.279 - 301

Received: 03 Apr 2024
Accepted: 25 Aug 2024
Published online: 05 Jan 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article