Modification of K-Means Clustering Algorithm for Optimizing School Zoning System Using Big Data

The increasing complexity of school zoning systems due to population growth and uneven geographic distribution has created significant challenges in ensuring balanced and fair allocation of students. This research focuses on modifying the K-Means clustering algorithm to optimize school zoning using Big Data. The admission of new students through the zoning system in Indonesia is crucial for enhancing access to quality education without discrimination. However, the current grouping process does not consider the distance between students' homes and schools. Clustering algorithms offer a solution by organizing data while accounting for the proximity between cluster members and their centroids. K-means clustering has shown promising results in grouping similar data but has limitations in determining centroid values, which can lead to suboptimal clustering. The modified K-Means algorithm is designed to address the limitations of traditional clustering methods in handling unbalanced group sizes, ensuring that each school zone has a proportional number of students. This study develops a clustering algorithm with fixed centroids to balance the number of members in each cluster. These algorithms are evaluated by calculating the total distance of each cluster member to its centroid. The proposed algorithms effectively cluster data and balance cluster membership, making them suitable for student placement in the zoning system. This research aims to provide a more effective solution for school zoning that can be applied to real-world educational systems, ultimately improving the fairness and efficiency of student distribution across schools.