A Novel Data Model Framework For Analytic In Online Education

Abstract

The rapid digitalization of education has created an urgent need for intelligent, scalable, and adaptable analytical systems. In response to this demand, this research introduces a novel integrated data model framework that effectively bridges demographic insights, device usage patterns, platform preferences, and learner feedback to support data-driven decision-making in online education. Unlike traditional models that focus solely on either engagement metrics or academic performance, the proposed framework offers a holistic understanding of learner behavior by analyzing a dataset comprising 1,000 responses with both structured and semi-structured attributes. At the core of this framework lies the Adaptive Randomized Fusion-layered Artificial NeuroNet (ARF-ANN), a custom deep learning architecture designed to integrate and enhance predictive capabilities. The multi-layered system includes stages for data preprocessing, encoding, and transformation through Python-based feature engineering pipelines. To ensure robustness and generalizability, the research also explores ensemble methods and hybrid modeling strategies. Results reveal significant correlations between profession types, device accessibility, and satisfaction ratings. Performance evaluation using metrics such as F1-score, precision, recall, and accuracy that ranged from 96% to 98% confirms the effectiveness of both ML and DL components. By embedding the innovative ARF-ANN and leveraging hybrid analytics, this framework marks a pioneering step toward real-time, personalized online learning analytics, enabling smarter and more adaptive educational ecosystems.