Estimating flood hazard rate in parepare using likelihood approach single decrement method

Abstract

Floods are one example of a random stochastic process. One important parameter to determine the chance of a flood to occur is the hazard rate. Therefore, a hazard rate estimation model is needed. One of the methods used to estimate the hazard rate at point t0 was the single decrement method with a likelihood approach that required exit time information, namely the time when a flood occurs and the assumed distribution of waiting times for the next flood to occur. The distribution of waiting times was assumed to be linear and exponential distribution. Hazard rate estimation used flood data that occurred in Parepare. The hazard rate estimator obtained using these two waiting time assumptions was transformed into a parametric model. The parametric model used was a regression model with linear, quadratic, and cubic assumptions. The best parametric model was a quadratic regression model for the assumed exponential distribution of waiting times based on R Square, Mean Square Error, and real regression test. The estimated hazard rate value obtained can be applied to estimate the probability of a flood event occurring in the interval (0,t0]. The selected parametric model is expected to be able to estimate the hazard rate value accurately.