# This situation is illustrated in figure 35 in this

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: recall that if is small, then fi (u) is the approximate probability that the observation is within a length interval centered at u, if Hi is the true hypothesis. For this reason, we still call fi (u) the likelihood of X = u if Hi is the true hypothesis. We also deﬁne the likelihood ratio, Λ(u), for u in the support of f1 or f0 , by f1 (u) Λ(u) = . f0 (u) A likelihood ratio test (LRT) with threshold τ is deﬁned by: Λ(X ) >τ <τ declare H1 is true declare H0 is true. Just as for the case of discrete-type observations, the maximum likelihood (ML) test is the LRT with threshold τ = 1, and the maximum a posteriori probability (MAP) decision rule is the LRT 114 CHAPTER 3. CONTINUOUS-TYPE RANDOM VARIABLES with threshold τ = π0 , where (π1 , π0 ) is the prior probability distribution. In particular, the ML π1 rule is the special case of the MAP rule for the uniform prior: π1 = π0 . The following deﬁnitions are the same as in the case of discrete-type observations: pfalse alarm = P (de...
View Full Document

## This note was uploaded on 02/09/2014 for the course ISYE 2027 taught by Professor Zahrn during the Spring '08 term at Georgia Institute of Technology.

Ask a homework question - tutors are online