Author: Eiko
Time: 2025-01-03 11:35:04 - 2025-01-03 12:34:16 (UTC)
Reference:
Theoretical Statistics by Robert W. Keener
Probability Theory by Eiko
Foundations of Modern Probability by Olav Kallenberg
Uniformly Most Powerful Tests
A test with level is called uniformly most powerful if for any , we have
This means we are not just optimizing for one nor for the power function, we are optimizing for all . It might not exist.
A family of densities , has monotone increasing likelihood ratios if there exists a statistic such that whenever , the likelihood ratio is a non-decreasing function of .
For the decreasing case it suffices to replace by .
Example. For densities of the form of exponential families (of single parameter)
if is non-decreasing / increasing, the likelihood ratio
is a non-decreasing / increasing function of . This includes the example of the exponential distribution, for which is increasing and .
Monotone Likelihood Gives Uniformly Most Powerful Tests
For this concept of monotone likelihood, we assume the parameter space is one-dimensional.
Suppose the family of density has monotone likelihood ratio (with respect to a test statistic ). Then
The monotone test
is a uniformly most powerful test for the hypothesis and , with level
and can be adjusted to achieve any desired level .
The power function for this test is strictly increasing when it is in , and non-decreasing overall.