Adaptive Tests
Abstract
The problem of hypothesis testing when the distribution is specified only up to a nuisance parameter is considered. A test is said to be adaptive if it is asymptotically optimal regardless of the value of the nuisance parameter. The exponential rate of convergence to zero of the probability of type II error when the probability of type I error converges to zero exponentially fast at a fixed rate is used as the optimal criterion. A necessary and sufficient condition for the existence of adaptive test is obtained.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1988
- Accession Number
- ADA199446
Entities
People
- K. S. Mak
Organizations
- Purdue University