Stock market prediction using Altruistic Dragonfly Algorithm.

Area of Science:

• Computational Finance
• Machine Learning
• Artificial Intelligence

Background:

• Accurate stock market prediction is crucial for financial decision-making.
• Traditional models often struggle with local minima and overfitting, limiting predictive performance.

Purpose of the Study:

• To propose and evaluate a novel hybrid model for stock market prediction.
• To enhance the predictive accuracy of Least Squares Support Vector Machine (LS-SVM) using an optimization algorithm.

Main Methods:

• A hybrid model combining the Altruistic Dragonfly Algorithm (ADA) with LS-SVM was developed.
• ADA was employed to optimize LS-SVM parameters, mitigating issues like local minima and overfitting.
• The model's performance was validated across 12 diverse financial datasets.

Main Results:

• The proposed ADA-LS-SVM model demonstrated superior predictive ability compared to other meta-heuristic algorithms.
• Experimental results confirmed the effectiveness of ADA in optimizing LS-SVM parameters for stock market forecasting.
• The hybrid model achieved better prediction performance, indicating its robustness.

Conclusions:

• The integration of ADA with LS-SVM offers a promising approach for more accurate stock market prediction.
• The study highlights the potential of meta-heuristic algorithms in refining machine learning models for financial applications.
• The proposed method provides a valuable tool for investors and financial analysts seeking reliable market insights.