Optimization Based Feature Selection and ML-Driven Anomaly Classification for Fog Environment

Debashis Dutta

doi:10.1002/itl2.70299

Optimization Based Feature Selection and ML-Driven Anomaly Classification for Fog Environment

2026-05-11

0

PRE

AI

DOI：10.1002/itl2.70299

Original

Original request for help

Save

Abstract

En 中文

Fog and edge computing environments, where resources are restricted, present significant security challenges due to the rapid growth of Internet of Things (IoT) applications. Traditional intrusion detection systems (IDSs) often fail to deliver high detection accuracy while maintaining low computational overhead in such decentralized settings. To address this limitation, this paper proposes an optimization-based feature selection and machine learning-driven anomaly classification framework tailored for fog environments. The proposed approach employs an Enhanced Seeker Optimization-based Feature Selection (ESO-FS) technique to identify the most relevant features, thereby reducing dimensionality and computational cost. A de-noising autoencoder is then utilized for effective anomaly detection, with its parameters optimally tuned using Particle Swarm Optimization (PSO) to enhance classification performance. Based on extensive experimental evaluations, the proposed ESO-FS model outperforms existing feature selection and intrusion detection techniques in terms of accuracy, loss convergence, and best cost metrics. According to the results, the proposed framework provides an efficient, lightweight, and scalable solution for intrusion detection in IoT fog and edge environments.

Keywords:

anomaly detection

feature selection

fog environment

internet of thing

optimization

Optimization Based Feature Selection and ML-Driven Anomaly Classification for Fog Environment

Abstract

Journal

Organization