MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202521099948 A) filed by Parul University Parul Institute Of Engineering And on October 16, 2025, for A System And Method For Anomaly Detection Utilizing Eigenvalue Decomposition In A Hybrid Computational Architecture For Financial Transactions And Iot Sensor Networks.
Inventors include Dr. Sanjay Agal; Dr. Mrudul Jani; and Dr. Payal Singh.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: A SYSTEM AND METHOD FOR ANOMALY DETECTION UTILIZING EIGENVALUE DECOMPOSITION IN A HYBRID COMPUTATIONAL ARCHITECTURE FOR FINANCIAL TRANSACTIONS AND IOT SENSOR NETWORKS The invention relates to anomaly detection in financial transactions and IoT sensor networks, leveraging a hybrid computational architecture. A Central Processing Unit first preprocesses data by performing noise filtration, normalization, and integrity verification before converting the information into a matrix format. The matrix data is then transmitted via a high-speed Data Bus to a dedicated Eigenvalue Decomposition Unit that computes eigenvalues and eigenvectors to capture dominant statistical features. The extracted eigenvalue-based features are stored and shared in a high-speed Random Access Memory before being processed by a Graphics Processing Unit executing parallel anomaly detection algorithms, including statistical analysis and machine learning classifiers. The overall system, interconnected by synchronous bidirectional data transfer, enhances the accuracy and efficiency of detecting anomalies in complex multi-source datasets.
Disclaimer: Curated by HT Syndication.
