|
Stock Price Prediction in Financial Markets Using a Hybrid GSO–SVR Algorithm
|
Javad Salari Askari1   , Hossein Shafiei *2 , Abbas Sheibani3 |
1- PhD student, Department of Accounting , si.c., Islamic Azad University, Sirjan, Iran
2- Assistant Professor, Department of Accounting , si.c., Islamic Azad University, Sirjan, Iran , hosseinshafiee@iau.ac.ir
3- Assistant Professor, Department of Accounting , si.c., Islamic Azad University, Sirjan, Iran |
|
|
Abstract: (8 Views) |
Purpose: This study examines the performance of a hybrid Golden Search Optimization–Support Vector Regression (GSO–SVR) model for predicting daily stock closing prices.
Methodology: This applied study adopts a quantitative approach. Daily closing-price data were collected for eight publicly traded companies: Apple, Google, Facebook, IBM, Microsoft, Twitter, Yahoo, and Ford. To preserve temporal order and reduce information leakage, the data were split chronologically, and hyperparameter selection was restricted to the validation set. The Golden Search Optimization algorithm was used to determine suitable values for the SVR penalty parameter and the radial basis function (RBF) kernel coefficient by minimizing the root mean square error (RMSE). The proposed model was compared with conventional SVR and genetic algorithm-optimized SVR (GA–SVR). Performance was evaluated using normalized mean squared error (NMSE), mean absolute error (MAE), and directional symmetry (DS).
Findings: The proposed GSO–SVR model achieved the lowest NMSE and MAE for all eight companies and the highest DS for seven of the eight companies. These results indicate improved numerical prediction accuracy and directional agreement compared with the benchmark models.
Conclusion: Hyperparameter optimization using the Golden Search Optimization algorithm may improve the forecasting performance of financial time series. However, its effectiveness should be evaluated empirically for each stock and performance metric.
|
|
| Keywords: Stock price prediction, support vector regression, Golden Search Optimization, genetic algorithm, hyperparameter optimization, financial time series. |
|
|
Full-Text [PDF 889 kb]
(4 Downloads)
|
Article type: Research |
Subject:
General
Received: 2026/07/25 | Accepted: 2026/10/26
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
