Design and Testing Novel One-Class Classifier Based on Polynomial Interpolation With Application to Networking Security

This work exploits the concept of one-class classifier applied to problem anomaly detection in communication networks. The article presents design an innovative algorithm based on polynomial interpolation technique and statistical analysis. method is datasets largely used scientific community for bench-marking such as KDD99, UNSW-NB15 CSE-CIC-IDS-2018, further evaluated with application a novel available dataset EDGE-IIOTSET 2022. paper also reports experimental results showing that proposed methodology outperforms classic classifiers, Extreme Learning Machine Support Vector models, rule-based intrusion system like SNORT. With respect binary this has advantage not requiring any a-priori knowledge about attacks collection only normal data traffic.