|
An Optimized GWO-LSTM Model for Forecasting the S&P 500 Index in the Stock Market
|
Arash Einy *1   , Mahnaz Khodaei2 |
1- Master of Accounting, Novin Non-Profit Organization, Ardabil , arash.einy1364@gmail.com
2- Master of Psychology, Azad University |
|
|
Abstract: (345 Views) |
Stock market forecasting plays a crucial role in financial decision-making; however, traditional models often struggle to capture the complex and nonlinear dynamics of financial data. Despite the widespread use of machine learning techniques to enhance prediction accuracy, many models still fall short due to suboptimal parameter tuning and limited adaptability to volatile market conditions. In this study, we propose a novel hybrid model called GWO-LSTM, which integrates the Grey Wolf Optimizer (GWO) with Long Short-Term Memory (LSTM) neural networks to optimize hyperparameters and improve stock price prediction accuracy. The model was tested using historical data of the S&P 500 index from 2015 to 2024, incorporating features such as open, high, low, volume, and closing prices. The model's performance was evaluated using three key metrics: the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). It was also benchmarked against baseline models including multilayer perceptron, artificial neural networks, and extreme learning machines. Furthermore, the Diebold-Mariano test and p-value analysis confirmed the statistically significant superiority of the proposed model. The findings indicate that GWO-LSTM can serve as a powerful tool for enhancing stock market forecasts, supporting financial decision-making, and improving risk management.
|
Article number: 10 |
| Keywords: Stock market prediction, machine learning, time series forecasting, financial management, risk management, GWO-LSTM, metaheuristic optimization |
|
|
Full-Text [PDF 2643 kb]
(83 Downloads)
|
Article type: Research |
Subject:
Special
Received: 2025/07/8 | Accepted: 2025/10/25 | Published: 2026/05/31
|
|
|
|
|
|
