Estimation of Generalized Pareto Distribution Parameters in Traffic Accident Loss Data Modeling

The problem of traffic accidents in Indonesia has a high level risk. In an effort to minimize losses due accidents, it is necessary study the data and characteristics identify these events as extreme events. This was conducted find out how estimate shape scale parameters using Maximum Likelihood Estimation (MLE), explore on accident Indonesia. method used analyze value Extreme Value Theory. One approach values Peaks Over Threshold which follows Generalized Pareto Distribution (GPD). Traffic loss divided into three types based cause, namely driver negligence, vehicle quality, other external factors period (2008-2017). obtained through MLE then solved by Newton Raphson because produces equations that are not closed form. resulted GPD distribution parameter, well confidence interval (1-α) 100% with 5%. addition, concluded from estimation have same for each type risk analyzed, but different parameter values. Based estimation, expected be useful related parties analyzing number next consider steps can taken reduce accidents.