MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112456 A) filed by Vellore Institute Of Technology on September 18, 2026, for A System For Early Parkinsonian Classification Using Waa-Based Dynamic Ensemble Selection And Method Thereof.
Inventors include Magesh; and Vinothini S.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: The present disclosure proposes a system (100) and method for early Parkinsonian classification using Weighted Average Accuracy (WAA)-based dynamic ensemble selection, patient-specific classifier-competence evaluation, reliability weighting, and decision fusion for improving classification robustness across heterogeneous voice, gait, and handwriting representations. The system (100) comprises signal-acquisition devices (108) for capturing biomarker signals corresponding to voice, gait, and handwriting, and a computing device (102) having modules (109) that preprocess signals, extract modality-specific features, generate a patient-specific representation, and maintain a heterogeneous classifier pool. An adaptive classifier weighting module (118) identifies a local competence region, determines WAA competence scores, selects classifiers, excludes unselected classifiers, and derives fusion weights from the same scores. An output interface module (122) provides Parkinsonian, non-Parkinsonian, or risk-probability classification data, and an explainability module (124) provides feature, classifier, fusion-weight, and confidence information.
Disclaimer: Curated by HT Syndication.
