MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641075663 A) filed by Sr University on June 18, 2026, for An Autoregressive Residual-Based Real-Time Drift Detection System For Financial Markets.
Inventors include Ankam Ramesh; and Rajchandar K.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: AN AUTOREGRESSIVE RESIDUAL-BASED REAL-TIME DRIFT DETECTION SYSTEM FOR FINANCIAL MARKETS The invention relates to an autoregressive residual-based real-time drift detection system for financial markets. The system employs rolling autoregressive models to predict price movements and calculates residuals to monitor deviations between predicted and actual prices. Residual volatility is tracked using a second rolling window, and a dynamic threshold based on moving averages and standard deviations is applied to detect abnormal market behavior. A drift score with exponential decay minimizes false positives, and alerts are generated when regime shifts are detected, transmitted to trading systems via API. The system operates without labeled data, is lightweight and interpretable, and can be deployed in real-time streaming environments using Kafka and Docker. By providing early warnings of regime shifts 18 to 36 hours before conventional methods, the invention enhances trading responsiveness, reduces drawdowns, and improves profitability in high-frequency financial markets.
Disclaimer: Curated by HT Syndication.
