MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202641115618 A) filed by Mr. K. Naupal Reddy; Dr. Smitha Mahindrakar; and Dr. M. Sandra Kirthy on September 26, 2026, for A Machine Learning-Based Fintech Platform For Real-Time Transaction Risk Scoring.
Inventors include Mr. K. Naupal Reddy; Dr. Smitha Mahindrakar; and Dr. M. Sandra Kirthy.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: The present invention discloses a machine-learning-based FinTech platform for real-time transaction risk scoring. The platform receives an electronic financial transaction request and generates a transaction feature representation comprising transaction, behavioural, temporal and contextual characteristics. A dynamic behaviour memory module maintains temporally differentiated behavioural representations comprising historical, recent and current-session transaction behaviour. A machine- learning inference engine processes the transaction feature representation and dynamic behavioural representations to generate a preliminary transaction risk value. An uncertainty estimation module determines an uncertainty value associated with the preliminary risk value. An adaptive calibration module generates a calibrated transaction risk score using the preliminary risk value, uncertainty value, behavioural deviation and transaction context. A transaction decision engine maps the calibrated risk score to a plurality of transaction-processing states comprising authorization, monitoring, additional authentication, temporary holding and blocking. Subsequent transaction outcomes are supplied to a feedback-learning module for updating the dynamic behavioural representation and/or machine-learning training information. The invention thereby provides an adaptive, uncertainty-aware and feedback-driven architecture for real-time transaction risk assessment and transaction processing in FinTech environments.
Disclaimer: Curated by HT Syndication.
