MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112170 A) filed by Vallurupalli Nageswara Rao Vignana Jyothi Institute Of Engineering And Technology on September 18, 2026, for A Machine Learning-Based Intelligent Security Framework For Real-Time Insider Threat Identification And Prevention.
Inventor includes Dr. K. Archana.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: ABSTRACT [0020] A machine-learning-based insider threat detection system is disclosed that continuously evaluates user activity logs to identify anomalous internal behavior. Three independent models Random Forest, Isolation Forest and Support Vector Machine—are trained on activity-derived feature sets that include login patterns, file access frequency, USB device usage and communication records. After preprocessing that scales numerical values and encodes categorical attributes, each model produces a threat probability or anomaly score. These scores are combined through a weighted ensemble to generate a single risk value for every monitored user. A web-based dashboard built with Streamlit displays the risk scores, visualizes behavioral trends and issues alerts when a predefined threshold is exceeded. Experimental comparison shows that Random Forest yields the highest overall accuracy, Isolation Forest effectively isolates rare outliers with few false positives, and Support Vector Machine maintains consistent performance across varied test conditions. The resulting platform therefore supplies administrators with timely, explainable insight into internal security risks while remaining scalable to large volumes of enterprise log data.
Disclaimer: Curated by HT Syndication.
