MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202641018269 A) filed by Kotha Raghavendra; Dr. Mamta Thakur; Mr M. Tirupathi Rao; Sudheer Nandi; Dr. Rajashekar Kandakatla; Dr. Srinivasan Nagaraj; Dr. Ramakrishnaiah; and Dayananda Sagar Academy Of Technology And Management, Kurnool, Andhra Pradesh, on Feb. 18, for 'system and method for real-time time series anomaly detection using machine learning and statistical process control.'
Inventor(s) include Kotha Raghavendra; Dr. Mamta Thakur; Mr M. Tirupathi Rao; Sudheer Nandi; Dr. Rajashekar Kandakatla; Dr. Srinivasan Nagaraj; and Dr. Ramakrishnaiah.
The application for the patent was published on Feb. 27, under issue no. 09/2026.
According to the abstract released by the Intellectual Property India: "This invention relates to a system and method for real-time time series anomaly detection using an integrated framework that combines machine learning techniques with statistical process control (SPC). The invention is designed to monitor continuous streams of time-dependent data generated from industrial systems, financial platforms, healthcare devices, telecommunications networks, energy grids, transportation systems, and other data-driven environments. By leveraging predictive modeling alongside statistically validated control mechanisms, the invention enables accurate and reliable detection of abnormal patterns, deviations, and emerging risks in real time. The system comprises a data ingestion and preprocessing module configured to collect high-volume and high-velocity streaming data from heterogeneous sources. A machine learning engine is employed to learn normal behavioral patterns using supervised, semi-supervised, or unsupervised models, including neural networks, autoencoders, support vector machines, and ensemble techniques. These models capture temporal dependencies, nonlinear relationships, seasonality, and multivariate interactions inherent in time series data. The learned models continuously adapt to evolving process conditions to address concept drift. A statistical process control module operates in conjunction with the machine learning engine to evaluate residual errors or anomaly scores using techniques such as control charts, cumulative sum (CUSUM), and exponentially weighted moving average (EWMA). Adaptive control limits are dynamically generated based on model outputs to ensure statistically significant validation of anomalies while minimizing false positives and false negatives. The invention further provides low-latency processing, automated model updating, scalable deployment across distributed computing environments, and enhanced interpretability through statistical metrics and visualization tools. By integrating adaptive learning with rigorous statistical validation, the invention delivers a robust, scalable, and intelligent solution for proactive monitoring and anomaly detection in real-time time series applications."
Disclaimer: Curated by HT Syndication.
