Pemodelan Kecelakaan Sepeda Motor Di Kota Surabaya Dengan Metode Glm (Generalized Linear Model)

The growth of motorcycles in the recent years is rapidly increasing, about16.54% per year. Due to this growth, the amount of accident on the roads hasincreased in number, where one of the areas that has increasing number ofaccident is Surabaya. This research aims to set up the prediction models ofmotorcycle accidents (MCA) and motorcycle victims (KSI).The method used is field research with the primary data based on thetraffic and the road geometric survey, while the data of accidents in the year 2006to April 2009 is as the secondary data which is obtained from the PoliceDepartment. The method used to analyze the data is Generalized Linear Models(GLM) with Poisson distribution and logarithm link function.From the model of motorcycle accident (MCA), it could be predicted thatthe increasing of either traffic or motorcycle proportion 10%, would increase thenumber of motorcycle accidents 37.86% and 42.78% per year. The increase of theroad width 1.0 meter is predicted to make the motorcycle accident decrease12.59% per year, while the road with the lines < 3 is predicted could decrease thenumber of motorcycle accident 44.46% per year. The model of motorcyclevictims who died or seriously injured (KSI) shows that the increase of traffic 10%would increase the motorcycle accident for the victims who died or seriouslyinjured 34.72% per year, while the limit of motorcycle proportion to < 70% andthe increase of road width 1.0 meter is predicted able to decrease the victims whodied or seriously injured 46.69% and 20.24% per year