3 Parametric Models
If we are willing to assume that survival times follow a particular family of distributions, such as the exponential or the Weibull, the whole survival curve is described by one or two parameters, which can be estimated by maximum likelihood. Parametric models give smooth estimates, allow extrapolation, and are efficient when the assumed family is right. The key step is to write down the likelihood correctly when some observations are censored. This section does that, fits the four most common families to the lung data, and shows how to choose among them.
3.1 The likelihood with censored data
3.2 The exponential model
3.3 The Weibull and other families
3.4 Choosing a model
3.5 Parametric models in software
3.6 Practice Problems
3.2 The exponential model
3.3 The Weibull and other families
3.4 Choosing a model
3.5 Parametric models in software
3.6 Practice Problems
Questions on this section
Stuck on something here? Ask below and it stays attached to this topic.