Hybrid Genetic Algorithm with Particle Swarm Optimization in the Fields of Artificial Intelligence and Data Sciences
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Artificial Intelligence (AI) and Data Science have become crucial to addressing complex issues across various domains. Since these are data-driven areas of inquiry, they often rely on large datasets consisting of many features, and the selection of features and model optimization play significant roles in predictive performance. However, conventional methods often face challenges such as computationally expensive features, lengthy convergence periods, and lower accuracy due to the presence of irrelevant or redundant features. To overcome these limitations, a novel hybrid Genetic Algorithm–Particle Swarm Optimization (GAPSO) was developed. The Genetic Algorithm (GA) performed the global search process for subsets of features and their combinations. In contrast, Particle Swarm Optimization (PSO) performed more local fine-tuning (probabilities of feature selection), which allowed for convergence towards the best solutions. The hybrid approach of GA-PSO has been evaluated on several well-known UCI datasets (Heart Disease, Iris, Wine, and Diabetes) using three classifiers (SVM, Random Forest, and Neural Networks). The results showed that the hybrid approach can reduce features by 25-46% and improve classification accuracy by 2-5% compared to either the GA-only method or PSO-only method. Overall, the evidence supports the idea that GA-PSO methods provide enhancements over AI and data science methods through predictive performance, computational efficiency, and model interpretability.
Fields of Study & Subjects
Metaheuristic Optimization Algorithms Research
Evolutionary Algorithms and Applications
Artificial Intelligence in Healthcare
Publication details
- Journal / source
- IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
- Pages
- 1298–1304
- Publication date
- 28 January 2026
- Language
- EN