Aproksimasi Numerik Belah Dua dan Newton-Raphson pada Estimasi Parameter Distribusi Weibul

Abstract

Life time distribution of warranted product, for example like automobile and photocopier is an interesting studies. Based on the fact that mostly, life distribution of warranted product is somewhat skewed to the right, this suggests that usual failure distribution such as Weibull distribution may provide a reasonable fit to the data. Related to these studies, parameter estimation is directly applicable since life distribution function is a function of parameters. There are no closed solution for Weibull likelihood function maximation. Therefore, numerical approximation could be an alternative solution. In this paper, Newton-Raphson and bisection method were employed to estimate the scale and shape parameter of Weibull distribution for photocopier failure data for since 4,5 years. Performance of those method were compared based on their approximation result and how fast they get into the solution. Data were taken from Bulmer and Eccleston (2003) research about photocopier realibility model.