Research On Analysis Of Chinese University Students' Learning Behaviors And Learning Outcomes Through Big Data Mining Visual Portrait As An Intermediary And Learning Early Warning As A Moderator

Abstract

This study explores the multifaceted factors influencing academic outcomes among Chinese university students through the lens of educational data analytics. Utilizing a mixed-methods approach, demographic characteristics, learning behaviors, early warning indicators, and the mediating and moderating effects of visual portraits and early warning systems are examined across three prominent Chinese universities: Tsinghua, Peking, and Fudan. Descriptive statistics reveal variations in mean age, gender ratios, and mean Grade Point Average (GPA) among students, while correlation and regression analyses highlight the significant positive associations between learning behaviors (such as study hours and online engagement) and academic outcomes (such as GPA). Chi-square tests demonstrate the predictive power of early warning indicators in identifying students at risk of academic underperformance. Additionally, mediation and moderation analyses elucidate the intermediary and moderating roles of visual portraits and early warning systems in shaping the relationship between learning behaviors and academic outcomes. Findings underscore the importance of adopting a holistic approach to student support and educational practice, informed by evidence-based interventions derived from educational data analytics.