Penerapan K-Means Clustering untuk Analisis Kondisi Lalu Lintas di Jalan Ir. H. Soekarno Surabaya

Abstract

Growth in the number of vehicles, especially in urban areas, has a significant impact on traffic density, especially during peak hours, so an approach is needed to group traffic conditions based on the volume of all types of vehicles and the degree of saturation using the K-Means Clustering algorithm. The data used are the volume of all types of vehicles and the degree of saturation obtained from the Surabaya City Transportation Agency. The clustering results show that there are 4 clusters of different traffic characteristics, such as the volume of 2-wheeled vehicles during heavy traffic conditions of more than 4700 vehicles with a degree of saturation of more than 0.45. Evaluation using the silhouette coefficient produces a value of 0.63, which means the quality of the cluster is in a medium structure. This study shows that the clustering method is effective in understanding traffic conditions, although additional features can be done to optimize the quality of the cluster.