Inference

CML includes two classes of methods for analyzing the distributions of the estimated parameters:

• confidence limits by inversion of the two types of chi-squared statistics - Wald and likelihood ratio
• simulation of the distribution of the parameters by weighted and unweighted bootstrap

The CML procedure for the weighted bootstrap implements a Bayesian simulation of the posterior distribution of the parameters based on Newton and Raftery's (1994) weighted likelihood bootstrap method.

• Confidence Limits by Inversion
• Bootstrap