MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641076771 A) filed by Vellore Institute Of Technology on June 21, 2026, for Adaptive Ensemble System For Imbalanced Network Intrusion Detection.
Inventors include Dr. P. Antony Seba; Steve Alex Mathew; Sankeerth Rajesh Nair; and Jesu Joel George.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: The present invention relates to an adaptive ensemble system (100) for imbalanced network intrusion detection. The adaptive ensemble system (100) comprises a preprocessing module configured to normalize network traffic features, a multi-stage feature selection module configured to combine global multi-class mutual information analysis and per-class binary mutual information analysis, a gradient-boosting classification module configured to process normalized feature representations, a principal component analysis module configured to generate orthogonal feature representations, and a weighted random forest classification module configured to generate weighted class probability outputs. The adaptive ensemble system (100) further comprises an adaptive optimisation module configured to determine optimized decision-tree weights using simulated annealing–based probability simplex optimisation associated with classification accuracy, false negative rate, recall variance, and minimum class recall parameters. The adaptive ensemble system (100) further serializes trained models, projection parameters, and optimized decision-tree weights for deployment-ready inference operations.
Disclaimer: Curated by HT Syndication.
