APPLICATION OF K-MEANS ALGORITHM IN KINDERGARTEN SCHOOL LOCATION CLUSTERING OF SCHOOL SELECTION STRATEGY BY PARENTS

This research aims to improve the kindergarten school location clustering model to support parents' school selection strategies. The main issue raised is the need to understand parents' preferences more deeply in choosing the right school for their children. To achieve this goal, the K-Means algorithm was applied and analyzed to cluster parents' data based on characteristics such as occupation, education, and residential location. This research utilizes a quantitative method with an exploratory descriptive approach. The results showed that the K-Means algorithm successfully formed two clusters with different characteristics. Cluster_0 includes groups with more centralized or close locations, education levels that tend to be low, and types of jobs that are at the lower middle economic level, while cluster_1 groups with more dispersed or distant locations, higher education levels, and jobs that are at higher economic levels. The quality of the resulting clusterization is considered quite good, with a Davies-Bouldin Index (DBI) value of 0.151. The application of the K-Means algorithm is proven to be effective in identifying groups of parents with different preferences, so it can be a foundation for schools in developing more targeted and tailored service strategies. This research makes an important contribution to the application of clustering techniques to support marketing strategies and decision-making in the early childhood education sector.