Article: Hyper-heuristic glowworm swarm optimised support vector machines for heart and thyroid disease classification Journal: International Journal of System of Systems Engineering (IJSSE) 2024 Vol.14 No.2 pp.226 - 246 Abstract: In order to improve illness detection accuracy and reduce complexity, this study seeks to construct sophisticated machine learning (ML) classifiers and effective feature selection (FS) methods. The proposed model includes two stages: classification using Hyper-heuristic Glow Worm Swarm Optimised Support Vector Machines and FS utilising Information Gain (IG) and Spotted Hyena Optimiser (IG-SHO) (HHGWSO-SVM). By removing the irrelevant characteristics with the IG metric, the dimensionality of attributes is decreased in the IG-SHO technique. By combining the hybrid optimisation approach of HHGWSO with the SVM, the suggested HHGWSO-SVM classifier has been created. Its configuration has been improved by optimally setting the margin parameter, kernel type, and kernel parameters. The Hyperheuristic algorithm and the Glowworm Swarm Optimisation (HHGWSO) have been combined to create a method for fine-tuning SVM parameters based on accuracy and model complexity. The proposed HHGWSO-SVM model is tested in experiments on benchmark datasets to predict thyroid and heart illnesses. According to the results, the suggested categorisation model has improved precision and accuracy while reducing model complexity. Inderscience Publishers - linking academia, business and industry through research

Title: Hyper-heuristic glowworm swarm optimised support vector machines for heart and thyroid disease classification

Authors: G. Kiruthiga; S.Mary Vennila

Addresses: Department of Computer Applications, Guru Nanak College, Chennai, 600 042, Tamil Nadu, India ' Department of Computer Science, Presidency College, Chennai, 600 005, Tamil Nadu, India

Abstract: In order to improve illness detection accuracy and reduce complexity, this study seeks to construct sophisticated machine learning (ML) classifiers and effective feature selection (FS) methods. The proposed model includes two stages: classification using Hyper-heuristic Glow Worm Swarm Optimised Support Vector Machines and FS utilising Information Gain (IG) and Spotted Hyena Optimiser (IG-SHO) (HHGWSO-SVM). By removing the irrelevant characteristics with the IG metric, the dimensionality of attributes is decreased in the IG-SHO technique. By combining the hybrid optimisation approach of HHGWSO with the SVM, the suggested HHGWSO-SVM classifier has been created. Its configuration has been improved by optimally setting the margin parameter, kernel type, and kernel parameters. The Hyperheuristic algorithm and the Glowworm Swarm Optimisation (HHGWSO) have been combined to create a method for fine-tuning SVM parameters based on accuracy and model complexity. The proposed HHGWSO-SVM model is tested in experiments on benchmark datasets to predict thyroid and heart illnesses. According to the results, the suggested categorisation model has improved precision and accuracy while reducing model complexity.

Keywords: heart disease; thyroid disease; machine learning; FS; information gain; spotted hyena optimiser; hyper-heuristic glowworm swarm optimisation; support vector machines.

DOI: 10.1504/IJSSE.2024.137067

International Journal of System of Systems Engineering, 2024 Vol.14 No.2, pp.226 - 246

Received: 07 Dec 2022
Accepted: 14 Feb 2023
Published online: 01 Mar 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Hyper-heuristic glowworm swarm optimised support vector machines for heart and thyroid disease classification

Keep up-to-date