K-Means Algorithm for Clustering the Location of Accident-Prone on the Highway

Abstract

In Indonesia, there is a highway which connects Jakarta and Bogor, some accidents have occurred at the highway every year. This paper aims to analyze the location of the accident to classify the high and low levels of accident vulnerability in the Jakarta Bogor Highway. First, the extracted data is then grouped based on some of the same characteristics in the dataset namely cause, location, minor injuries, serious injuries and death. Second, the grouping results are visualized in the form of highway maps that can help highway managers in identifying and evaluating several accident-prone points on Jakarta Bogor Highway. The method to be used in data processing in this study is the K-Means clustering algorithm which is expected to produce useful information for Jakarta Bogor Highway managers. The results of this study indicate that accidents that often occur are in cluster 3 with a total of 80 accidents and at least there are in cluster 2 with a total of 57 accidents, the location of accidents that often occurs in cluster 1 is in KM 24, while cluster 2 is in KM 41 and cluster 3 located at KM 10.6.