<![CDATA[Inclusive education requires integrated academic, social, and emotional support systems to address learner diversity; however, existing Educational Data Mining (EDM) research has largely concentrated on academic or log-based data, with limited attention to family-centered dynamics. This study aims to develop and validate a family-centered EDM framework to examine how the quality of parent–child interactions predicts students’ reflective learning behaviors. A cross-sectional survey was conducted with 150 junior and senior high school students from urban Indonesian schools using the Parent–Child Inclusivity Scale (PCIS) and the Self-Assessment Behavior Inventory (SABI). Data were analyzed using descriptive statistics, reliability testing, and Confirmatory Factor Analysis, followed by EDM techniques including Decision Tree and Naïve Bayes classification, K-Means clustering, and Apriori association rule mining. The classification models achieved accuracies of 82% and 78%, while clustering revealed three distinct inclusivity–reflection profiles. Association rules further identified key combinations of autonomy support, constructive feedback, and joint decision-making that consistently predict high levels of reflective behavior. By explicitly integrating family-centered variables into predictive, profiling, and rule-based EDM analyses, this study advances Informatics by proposing a novel family-centered EDM framework that bridges data analytics with inclusive educational practice and adaptive learning system design.]]>
