A Bayesian Approach to the Classification Problem
- 877 Downloads
When measurements are present/absent and when the prevalence p of the individual samples possessing the trait is unknown, the optimal composite sample size can be determined for every retesting procedure. Chapter 2 contains the discussion and derivations of the results necessary for this purpose. It is clear form these derivations that the optimal composite sample size depends on the prevalence p. Since the prevalence is usually not known prior to sampling, it may not be possible to determine the optimal composite sample size. A Bayesian approach is suggested in such situations.