Minimax-Optimal Strength of Statistical Evidence for a Composite Alternative Hypothesis

David R. Bickel
  • International Statistical Review, July 2013, Wiley
  • DOI: 10.1111/insr.12008

Criteria for a measure of evidence and two optimal measures of evidence

What is it about?

How strongly do observations support a hypothesis? This paper gives general criteria to be satisfied by any measure of the strength of evidence. It also defines two ways such a measure might be considered optimal.

Why is it important?

While the likelihood ratio measures the strength of evidence of one distribution over another, there is no general agreement on how to measure the strength of evidence for more general hypotheses. This paper explores some ways to do that using advances in information theory.

Perspectives

David R. Bickel
University of Ottawa

This publication generalizes the likelihood measure of evidential support for a hypothesis with the help of tools originally developed by information theorists for minimizing the number of letters in a message. The approach is illustrated with an application to proteomics data.

Read Publication

http://dx.doi.org/10.1111/insr.12008

The following have contributed to this page: David R. Bickel