Modeling factor as the cause of traffic accident losses using multiple linear regression approach and generalized linear models

Abstract

Road Traffic and transport as part of the national transportation system that should be developed its potential and role to realize security, safety, order, and smoothness of traffic and Road Transport to support economic development and regional development. Traffic accidents in Indonesia are currently ranked second in the ASEAN region, the number of traffic accidents an average of 28,000 to 30,000 people per year. Large casualties cause high material losses; this can result in poverty levels. The level of poverty experienced by traffic accidents due to natural disadvantages requires care, lost productivity, lost livelihood, stress and prolonged suffering. This study aims to model the factors causing the magnitude of traffic accidents using linear regression model and the GLM model. The results obtained by factor is out of balance and exceeds the speed limit which has a significant effect on the amount of traffic accident loss for both model linear regression model and GLM model, while the disorder factor and the influence factor of alcohol have no significant effect. Multiple linear regression model obtained can be written in the form of equations i.e. Y = 2367906 + 10716421 X 1 + 987344 X 4, and the GLM model equations can be written i.e. ln(μ) = 17,73500 + 0,05264X 1 + 0,09291X 4 From both models obtained the best model based on model which has the smallest AIC value that is GLM model.