MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641093063 A) filed by Elizabeth Antony; and Amit Kumar Roy on July 31, 2026, for An Enhanced Intrusion Detection System Using Stability-Aware Genetic Algorithm-Based Feature Selection With A Persistent Global Feature Memory.
Inventors include Elizabeth Antony; and Amit Kumar Roy.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present invention relates to a computer implemented system and method for enhancing an Intrusion Detection System (IDS) using stability-aware genetic algorithm (GA) based feature selection integrated with a persistent Global Feature Memory (GFM). Network traffic data are preprocessed and encoded into candidate feature subsets, which are optimized through a GA based optimization process employing selection, crossover, mutation, and elitism. A memory-guided composite fitness function evaluates each candidate feature subset based on intrusion detection performance, feature subset reduction, and historical feature selection information derived from the GFM. The GFM is progressively updated across multiple independent optimization runs to preserve feature selection knowledge and guide subsequent optimization. A stability assessment identifies a stable feature subset for use by the intrusion detection engine to detect malicious network activities, including botnet attacks. The disclosed system improves the stability, reproducibility, and computational efficiency of feature selection while enhancing the performance and reliability of machine learning based IDSs.
Disclaimer: Curated by HT Syndication.
