مجلة العلوم النفسية والتربوية

The Impact Of Using A Supervised Learning System On Student Performance In A Humanities And Social Sciences Course

Auteurs : Bahloul Farouk .

Résumé

This study examines the impact of integrating a locally hosted supervised learning system into the teaching of a data analysis module within an Information and Communication Sciences undergraduate programme at the University of Béjaïa. Adopting a descriptive, design-based methodology, the research targets the validation of an adaptive pedagogical instrument across one lecture group and six tutorial groups per cohort. Student performance is operationalized as a process-based, multi-dimensional construct monitored through five interaction indicators — reading time, GSE query logs, chatbot interactions, MCQ scores, and graded exercise results — aggregated via a backend dashboard. The system's architecture draws on Self-Regulated Learning theory (Zimmerman, 2000), Vygotsky's Zone of Proximal Development (1978). Preliminary correlational analysis reveals a positive association between MCQ performance and the conceptual specificity of student queries, while passive reading time alone does not predict academic outcomes. These findings suggest that AI-driven adaptive environments measurably improve students' cognitive engagement and self-regulatory behavior in heterogeneous learner contexts.

Mots clés

adaptive learning systems; supervised machine learning; process-based assessment; self-regulated learning; natural language processing; higher education.