ABSTRACT
This chapter focuses on methods based on the likelihood function. It shows that both parts of the model are determined by a limited number of unknown parameters. The chapter examines briefly 'exact' confidence intervals and tests. The simplest special case is the two-sample problem with interest concentrated on the ratio of the rate parameters. The exponentially based model with fixed explanatory variables is a special case of the accelerated life model. If that model is to be used with censored data, it is helpful if both density and survivor function can be computed in reasonably explicit form. The asymptotic relative efficiency of least squares relative to maximum likelihood can be calculated. For examining the dependency on explanatory variables, whether new or already in the model, direct plotting can be used with distinguished marking of the censored values.
