Negative Selection Algorithm in Static and Dynamic Data Environments: A Performance Analysis

Rejeesh E, Dr. Shijo M Joseph, Anupama M

Abstract: Network Intrusion Detection Systems (NIDS) increasingly rely on bio-inspired computational paradigms to cope with the volume, velocity, and variety of modern traffic. Among these, the Negative Selection Algorithm (NSA), a core mechanism of Artificial Immune Systems (AIS) modelled on self/non-self discrimination in the mammalian immune system, has been extensively investigated for anomaly-based intrusion detection because it requires only normal (self) samples for training and can, in principle, generalize to previously unseen (non-self) attacks. This paper reviews the evolution of NSA from its original binary, r-contiguous-bit formulation to real-valued, variable-sized-detector, density-guided, and hybrid deep-learning-assisted variants. The performance of these variants is examined separately for static benchmark datasets (e.g., NSL-KDD, KDD-Cup99, UNSW-NB15, CICIDS) and for dynamic, high-velocity, and concept-drifting environments typified by Internet of Things (IoT) and Industrial IoT networks. This paper is a comparative analysis of the performance of NSA variants and identified open challenges: detector-hole formation, curse of dimensionality, computational cost of detector regeneration, and the absence of native mechanisms for continual/incremental learning are discussed. This generated an insight to advancement of NSA to cover a decisive gap remains in ability to self-adapt to non-stationary, rapidly growing IoT data streams and lightweight NSA architectures suited to edge deployment.

Keywords: Negative Selection Algorithm, Artificial Immune System, Intrusion Detection System, Anomaly Detection, Internet of Things, Concept Drift, Dynamic Data Streams, NSL-KDD

How to Cite?: Rejeesh E, Dr. Shijo M Joseph, Anupama M, "Negative Selection Algorithm in Static and Dynamic Data Environments: A Performance Analysis", Volume 15 Issue 8, August 2026, International Journal of Science and Research (IJSR), Pages: 401-404, https://www.ijsr.net/getabstract.php?paperid=SR26731191252, DOI: https://dx.doi.org/10.21275/SR26731191252

Download Citation: APA | MLA | BibTeX | EndNote | RefMan