Related Experiment Video
Updated: Nov 7, 2025
09:17
Surrogate Model Development for Digital Experiments in Welding
Published on: March 28, 2025
1.4K
Forecasting stock price using integrated artificial neural network and metaheuristic algorithms compared to time
Milad Shahvaroughi Farahani1, Seyed Hossein Razavi Hajiagha2
1Department of Finance, Faculty of Management and Finance, Khatam University, Tehran, Iran.
Soft Computing
|May 3, 2021
Summary
This study predicts stock market indices using artificial neural networks (ANN) trained with metaheuristic algorithms like social spider optimization (SSO) and bat algorithm (BA). Results were compared against traditional time series models for enhanced stock price prediction.
Area of Science:
- * Computational Finance
- * Machine Learning
- * Econometrics
Background:
- * Stock markets serve as key economic indicators, offering investment opportunities but requiring sophisticated prediction methods.
- * Accurately forecasting stock price movements is crucial for investors and economic analysis.
- * Traditional time series models have limitations in capturing complex market dynamics.
Purpose of the Study:
- * To predict stock price indices using Artificial Neural Networks (ANN) enhanced by metaheuristic algorithms.
- * To evaluate the efficacy of Social Spider Optimization (SSO) and Bat Algorithm (BA) in training ANNs for stock prediction.
- * To compare the performance of ANN-metaheuristic models against established time series models like ARMA and ARIMA.
Main Methods:
- * Feature selection using Genetic Algorithms (GA) to identify relevant technical indicators.
- * Training ANNs with metaheuristic algorithms (SSO, BA) for stock index prediction.
- * Employing Mean Absolute Error (MAE) as a primary error evaluation metric.
- * Comparative analysis against Autoregressive Moving Average (ARMA) and Autoregressive Integrated Moving Average (ARIMA) models.
Main Results:
- * ANN models trained with SSO and BA demonstrated superior performance in predicting stock price indices compared to traditional time series models.
- * Genetic Algorithms effectively identified key technical indicators, improving model accuracy.
- * The study provides empirical evidence for the effectiveness of metaheuristic-optimized ANNs in financial forecasting.
Conclusions:
- * Metaheuristic-optimized Artificial Neural Networks offer a powerful approach for accurate stock market index prediction.
- * Integrating advanced machine learning techniques with feature selection enhances predictive capabilities in financial markets.
- * The findings suggest a promising direction for developing more robust stock forecasting tools.
