An Innovative Hybrid Approach for Efficient Anomaly Detection with Machine Learning and Optimization

EasyChair Preprint 15623

Abstract

Anomaly detection is a pivotal task across domains such as cybersecurity, finance, and healthcare. While machine learning has advanced significantly, designing algorithms that achieve an optimal balance between computational efficiency and detection accuracy continues to pose challenges. This paper presents a novel hybrid approach integrating Particle Swarm Optimization (PSO) with a Neural Network (NN) to enhance the effectiveness of anomaly detection. PSO is employed for feature selection and hyperparameter tuning, while the NN provides reliable classification capabilities. Experimental evaluations on benchmark datasets reveal notable enhancements in both accuracy and computational efficiency compared to state-of-the-art methods.

Keyphrases: Particle Swarm Optimization, anomaly detection, hybrid algorithm, machine learning, neural networks

@booklet{EasyChair:15623,
  author    = {Piter Wen and Li Wei and Mo Zhang and Hoo Wang and Michael Lornwood},
  title     = {An Innovative Hybrid Approach for Efficient Anomaly Detection with Machine Learning and Optimization},
  howpublished = {EasyChair Preprint 15623},
  year      = {EasyChair, 2024}}