8 Likelihood
Likelihood connects observed data to the parameters of a statistical model. It sits behind parameter estimation, profile constructions, and many common test statistics.
At the scaffold level it is enough to keep three points in view:
- likelihood is treated as a function of parameters for fixed data;
- the log-likelihood is usually the more practical object;
- the local curvature near the maximum is tied to the precision of the estimate.
8.0.1 Companion notebook
The chapter notebook builds a simple binomial log-likelihood and shows what its one-parameter profile looks like.