Abstract
<title>Abstract</title> <p>This study examines the effectiveness of circular economy implementation within Education for Sustainable Development (ESD) using a combination of statistical analysis and machine learning approaches. The research involved 505 respondents from educators, students, administrators, and policymakers across Jakarta and Banten. Data were analyzed using descriptive statistics, correlation analysis, regression and classification models, and feature importance techniques to understand the relationships among institutional, pedagogical, and policy-related factors. The findings indicate that perceived effectiveness of circular economy integration tends to be at a moderate level, with no extreme variation across respondents. Multiple determinants, including teacher competence, institutional support, curriculum integration, policy alignment, student engagement, and technological support, are interrelated and jointly associated with effectiveness outcomes. However, their relative contributions vary, suggesting a distributed pattern of influence rather than a single dominant factor. Machine learning results show relatively comparable performance across models, with ensemble-based methods providing slightly more stable predictive results, although overall predictive capacity remains moderate. The study highlights that ESD effectiveness in circular economy implementation is shaped by interconnected educational and institutional dimensions that require integrated attention. These findings provide empirical insights for strengthening sustainability education practices and demonstrate the potential of data-driven approaches in evaluating complex educational programs.</p>