English

Abstract

School participation is measured by the Pure Participation Rate (APM). This study examines whether data mining can generate new knowledge. The Central Sumatra Statistic Central Agency (BPS-North Sumatra) provided secondary statistics on APM by city/district (2011–2019) for elementary, junior high, high school, and PT. Data mining uses clustering (k-means) and classification (Decision tree). This cluster maps the APM. Mapping clusters are utilized again for categorization. Cluster value ranges indicate classification. C1 was the high APM cluster, and C2 was the low APM cluster. RapidMiner aids processing. The study found 18 high-cluster (C1) cities and 15 low-cluster cities (C2). Based on the clustering results obtained, classification results show that SMA and PT become influential attributes in mapping the area based on the Decision tree method, resulting in 3 rules: if SMA has a percentage 68,085% and PT has a presentation 18,730%. (high cluster). Classification and clustering have yielded new data.