MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641051632 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on April 23, for 'adaptive modular neural network system for iot botnet detection using stacking fusion.'
Inventor(s) include Asis Kumar Tripathy; and Ganduri Suhrud Sharma.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a system (100) and method for detecting botnet activities in Internet-of-Things networks using flow-level network traffic features. The system (100) comprises an input module for receiving network traffic data, a preprocessing module for data transformation, and a feature engineering module for generating behavioral features. A modular expert analysis module includes a plurality of heterogeneous expert models configured to generate probabilistic outputs based on the processed features. An adaptive fusion module combines the probabilistic outputs using a stacking-based meta-classifier to produce a fused output. A classification module generates a final output based on the fused output. The system (100) further includes an evaluation module for assessing performance using predefined metrics. The modular architecture and stacking-based fusion enable structured processing of network traffic data using multiple expert models within a unified framework."
Disclaimer: Curated by HT Syndication.
