MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202611001007 A) filed by Indian Institute Of Technology Banaras Hindu University, Varanasi on January 05, 2026, for System And Method For Automatic Threshold Determination In Extreme- Value Statistical Analysis.
Inventors include Pandey, Kamlesh Kumar; and Dixit, Shubham.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: The disclosure relates to a computer-implemented analytical system (100) and method (300) for automated identification of statistically significant thresholds in Extreme-Value-Theory (EVT) based analyses. The system (100) includes a server (102) equipped with one or more processors (104) and a memory (106) storing executable instructions configured to receive time-series datasets from a client device (110) and/or local storage, perform data pre-processing, and implement Generalized Pareto Distribution (GPD) fitting to exceedances above multiple candidate thresholds. An exponential limiting-case model is further evaluated for comparison. Diagnostic outputs such as parameter-stability plots, return-level plots, agreement plots, and information-criterion outputs (e.g., AIC/AICc) are generated to evaluate model behaviour and convergence across candidate thresholds. A rule-based threshold selection module (220) analyzes quantitative indicators and automatically identifies a statistically consistent threshold representing dominant tail behaviour. The system (100) provides processed datasets, analytical summaries, and diagnostic visualizations to the client device (110). The disclosed system (100) enhances accuracy, objectivity, and reproducibility in EVT-based threshold determination while reducing manual effort and improving analytical efficiency.
Disclaimer: Curated by HT Syndication.
