Identification of Traffic Accidents Vulnerability Level Using Kernel Density And K-Medoids Methods (Case Study: Depok and Kalasan Districts, Sleman Regency)

Abstract

One of the analytical tools that can be used to help, parse, identify and map traffic accident problems in an area is a Geographic Information System (GIS). GIS is used to create clusters of traffic accident events. The level of accident vulnerability in this paper is obtained by calculating the density of the number of incident points where the accident occurred, namely Depok and Kalasan Districts, Sleman Regency on a road segment length of 1,000 m per year. The clustering methods used are kernel density and k-medoids methods. Comparison of the identification of traffic accident-prone levels in Depok and Kalasan sub-districts using the Kernel Density and K-Medoids methods with a road length of 1,000 meters using the Kernel Density and K-Medoids methods in 2018 there is the same difference, namely 6.17% with the medium level classification and low level classification. For 2019 it is 0.01% with a high level classification, 1.85% for the medium level and 1.86% for the low level. For 2020 there is the same difference, namely 1.23% with medium and low level classifications. For 2021 there is no difference for high level classification but there is the same difference for medium and low level classification which is 3.7%.