PEMODELAN PREDIKSI KECELAKAAN LALU LINTAS PADA JALAN NASIONAL DI KOTA BANDA ACEH DITINJAU DARI FAKTOR LALU LINTAS DAN GEOMETRIK JALAN

Abstract

The population of Banda Aceh has been rapidly increasing in recent years, which has increased traffic and raised the number of accidents on the road. This study aims to build road accident models based on some factors using General Linear Model (GLM) method with Poisson regression. The primary data consisted of traffic volume, traffic speed, and road geometry. The secondary data were obtained from the Banda Aceh City Police Department, which included information on traffic accidents that occurred on national roads between 2018 and 2020. Due to its comparatively high accident risk, the National Roads in Banda Aceh City were chosen as the research location. The prediction model analyzed in this study was the number of accidents. The findings indicated that the number of accidents was positively correlated with traffic volume, the presence of u-turns, and the existence of intersections. The total lane width, meanwhile, had a negative impact on the number of accidents. Based on modeling results, it is forecasted that a 10% increase in traffic volume will lead to an increase in accidents by 1.021 times. Additionally, it is predicted that the number of accidents at the observation sites with the U-turn will increase by 2.83 times. While it is predicted to increase the number of accidents by 1.55 times per year at the observation point which has more than one intersections. Further, the 1 m addition of lane width is predicted to reduce the number of accidents by 1.72 times per year.