Year: 2024
Type: article
Abstract: Artificial Immune Recognition System (AIRS) offers a promising meta-heuristic approach inspired by the human immune system for classification tasks. However, limitations such as reliance on single-antigen activation and retention of untested memory cells can lead to inaccuracies. This paper proposes the Optimized Artificial Immune Recognition System (O-AIRS) to mitigate these issues. O-AIRS leverages Homogeneous Antigen Groups (HAGs) for refined memory cell activation, ensuring a precise threat response. Furthermore, O-AIRS incorporates a robust maturity mechanism to retain only validated memory cells, enhancing classification accuracy. The effectiveness of O-AIRS was assessed using established medical datasets: Liver Disorders (LD) and Haberman Surgery Survival (HSS). Experimental evaluation on both LD and HSS datasets establishes O-AIRS's superiority over AIRS and AIRS2 across various performance metrics. Notably, O-AIRS achieves this enhanced performance while utilizing approximately 50% fewer memory cells during classification due to its optimized activation mechanism. Importantly, O-AIRS guarantees the maturity of all memory cells, ensuring effective threat recognition.
Source: Informatica
Authors: Merad Badr, Belkadi Khaled
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4403431272