Analysis of Different Variants of Relief Based Algorithms for Feature Selection in Medical Applications

• Published in: 2023 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS) pp. 893 - 899
• Main Authors: Kumar, Alok, Choudhry, Mahipal Singh
• Format: Conference Proceeding
• Language: English
• Published: IEEE 03.11.2023
• Subjects: Classification algorithms; Complexity; Data analysis; Feature extraction; Feature Selection; Feature weight; Filter; Filtering algorithms; Lung cancer; Medical services; Relief; Task analysis
• DOI: 10.1109/ICCCIS60361.2023.10425149

This research paper explores the efficacy of Relief-based algorithms (RBAs) in the domain of medical data analysis, with a specific focus on lung cancer classification. Faced with the growing complexity of biological data and the need for computationally efficient yet sophisticated feature selection methods, RBAs emerge as a compelling solution. Various RBA variants, including ReliefF, Iterative Relief, I-Relief, SURF and MultiSURF, offer distinct perspectives on feature weighting and selection. Empirical analysis on lung cancer datasets reveals notable results, highlighting the impressive 0.80 accuracy rate achieved by I-Relief in classification tasks, thus confirming its effectiveness. These findings contribute to a deeper understanding of RBA capabilities and provide valuable insights for their future application and refinement within the domain of medical data analysis.