Censored data is often an uncontrolled reality in many data sets. Additionally, an incorrect estimation of the distribution parameters is a common side effect if this is ignored. Maximum Likelihood Estimation (MLE) of the parameters of censored distributions provides very accurate parameter estimators.
In teaching students to first understand the steps involved in obtaining estimated parameters for censored data, they will become more prepared to accurately deal with it in general data analysis. While tools exist in order to handle censored datasets, learning the fundamentals of estimation of censored distributions is important. My focus lies in understanding how and why using MLE will more accurately estimate distribution parameters for censored datasets, as a result providing a template through which students will better learn the necessary methods to deal with censored data.
I present four suggested steps to better understand the estimation of parameters when dealing with censored data. Understanding the concepts takes place through application of mathematical statistics, simulation, working with statistical software, and through analysis of an actual data set.
|