MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641092315 A) filed by Vardhaman College Of Engineering on July 30, 2026, for A System And Method For Machine Learning-Based Predictive Data Analytics.
Inventors include Mr. Naresh Goud M; Ms. Shalini K; Dr. Jagini Naga Padmaja; Mr. M Raghavendra; Dr. Karnati Ramesh; and Dr. Raghavendra Gowda D.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: ABSTRACT A System and Method for Machine Learning-Based Predictive Data Analytics The present disclosure relates to a system and method for predictive data analytics utilizing machine learning. The system receives operational data from multiple sources with time stamps and synchronizes the data in time. A provenance and integrity scoring module assesses the reliability of data pieces based on source identity, completeness, communication condition, duplication indication, and historical source performance. A causal feature resolution engine selects predictive features with and stable impact across different operating windows. A predictive inference engine generates a predicted technical condition and a confidence envelope generator derives prediction certainty based on source reliability, feature stability, data sufficiency and inference consistency. A drift-and-anomaly separation module classifies deviations into operational anomalies, transient noise, source faults, and predictive-model drift. Without confidence, an adaptive data-acquisition controller selectively requests high-information additional data and re-evaluates the prediction. It reduces false alerts, bandwidth consumption, processing and model updates to generate a validated alert, corrective-action recommendation, or control signal.
Disclaimer: Curated by HT Syndication.
