MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202641062229 A) filed by Ms. Pragathi Vulpala; Ms. Vasannagari Pavani; Ms. Prasanna Pasunari; Ms. K Srujana; Mr. Jagdish Kumar; and Mr. B. Srikanth, Hyderabad, Telangana, on May 16, for 'voting classifier based network intrusion detection system for communication networks.'
Inventor(s) include Ms. V. Sravanthi; Ms. V. Pranavi; Mr. V. Akhilesh; Mr. T. Sainath; Ms. Pragathi Vulpala; Ms. Vasannagari Pavani; Ms. Prasanna Pasunari; Ms. K Srujana; Mr. Jagdish Kumar; and Mr. B. Srikanth.
The application for the patent was published on May 29, under issue no. 22/2026.
According to the abstract released by the Intellectual Property India: "A computer-implemented network intrusion detection system is disclosed in which preprocessed network-flow features are supplied to a heterogeneous ensemble comprising Extra Trees, HistGradientBoosting, XGBoost, and Random Forest learners. A feature- conditioning stage applies a predetermined ranked subset, such as a 40-feature subset obtained from CHI-REV or another ranking method, and a soft-voting stage averages class- probability outputs from the learners to generate a final intrusion label and attack indicator. The system is deployable through an authenticated web interface and API that reuse stored scaling bounds, selected-feature order, and label mapping at inference time. In repository measurements on CICIDS-2017, voting embodiments improved selected baseline metrics, including all-class accuracy and macro-F1, while preserving compatibility with live analyst workflows."
Disclaimer: Curated by HT Syndication.
