MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202641123282 A) filed by S. Hrushikesava Raju; Suneetha Bulla; and Koneru Lakshmaiah Education Foundation on October 14, 2026, for Socioacad-Net: A Cluster-Aware, Attention-Fusion Ensemble System And Method For Predicting Student Academic Performance From Digital And Social Media Behavioural Data.
Inventors include Addepalli S V S Deeraj; Cheebrolu Nuthan Sai; Jasti Nithin Chowdary; Nukavarapu Ganesh Sai; and Dr. Suneetha Bulla.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: A computer-implemented system and method are disclosed for predicting student academic performance by fusing academic-record data with engineered digital and social-media behavioural features. A pre-processing and feature-engineering module computes a composite "Social Procrastination and Sentiment Volatility Index" (SPSVI) from late-night activity, sentiment volatility, pre-deadline activity spikes and academic-content ratio. A Digital Behaviour Persona Clustering (DBPC) module assigns each student to a latent behavioural persona using unsupervised clustering. A Temporal Engagement Attention (TEA) module represents each student's weekly engagement as a sequence, processed by a GRU network and a self-attention layer to yield a fixed-length engagement embedding. A Cluster-Conditioned Stacking Ensemble (CCSE) module combines a plurality of base learners and the engagement embedding through a meta-learner whose weighting is conditioned on the assigned persona, generating a performance class or risk score. An explainability module computes SHAP-based feature attributions and issues an early-warning alert when the risk score exceeds a threshold, enabling timely, explainable and subgroup-aware academic intervention.
Disclaimer: Curated by HT Syndication.
