MUMBAI, India, Sept. 7 -- Intellectual Property India has published a patent application (202641104236 A) filed by Vardhaman College Of Engineering on August 29, 2026, for System And Method For Ai-Based Student Performance Prediction.
Inventors include Dr. Swati Sucharita; Ms. Budavarapu Pravallika; Mr. Allama Prabhu; Ms. Vijaylaxmi; Mr. Majjari Sudhakar; and Mr. Sai Madhav Raj Alakuntla.
The application for the patent was published on September 04, 2026, under issue no. 36/2026.
Abstract: ABSTRACT SYSTEM AND METHOD FOR AI-BASED STUDENT PERFORMANCE PREDICTION The present disclosure relates to a system and method for AI- based student performance prediction using reliability-qualified educational data. Heterogeneous events from learning-management, assessment, attendance and related institutional sources are mapped to canonical learner, course and time references and assigned event-level or window-level reliability values according to completeness, recency, consistency and validity. A temporal state builder generates a reliability-qualified student-state representation that is processed by an artificial-intelligence prediction engine to produce one or more future performance predictions. Model uncertainty, data uncertainty and distribution drift are evaluated to obtain calibrated prediction confidence. A confidence gate selectively releases, qualifies, reprocesses or suppresses a prediction and can trigger source re-query or additional evidence acquisition when confidence is insufficient. Subsequently validated educational outcomes are reconciled with prior predictions for calibration, drift detection and validation-gated model updating. The system provides auditable educator-facing decision support while distinguishing missing evidence from negative performance evidence.
Disclaimer: Curated by HT Syndication.
