Let f0(x) be the p.f. of the Bernoulli distribution with parameter 0.3, and let f1(x) be the p.f. of the Bernoulli distribution with parameter 0.6. Suppose that a single observation X is taken from a distribution for which the p.d.f. f (x) is either f0(x) or f1(x), and the following simple hypotheses are to be tested: H0: f (x) = f0(x), H1: f (x) = f1(x). Find the test procedure for wh
Read moreTable of Contents
Textbook Solutions for Probability and Statistics
Question
Suppose that a single observation X is taken from the uniform distribution on the interval \([0,\ \theta]\), where the value of \(\theta\) is unknown, and the following simple hypotheses are to be tested:
\(\begin{array}{ll}H_0:&\theta=1,\\ H_1:&\theta=2\end{array}\)
a. Show that there exists a test procedure for which \(\alpha(\delta)=0\) and \(\beta(\delta)<1\)
b. Among all test procedures for which \(\alpha(\delta)=0\), find the one for which \(\beta(\delta)\) is a minimum.
Solution
Step 1 of 3
(a) Consider the following test \(\delta\) which rejects the null if and only if X > 1. Then,
\(\alpha(\delta)=P_{H_0}(X>1)=P(X>1\mid X\sim U(0,\ 1))=0\)
and
\(\beta(\delta)=P_{H_1}(X>1)=P(X>1\mid X\sim U(0,\ 2))=\frac{1}{2}<1\)
Thus this test \(\delta\) satisfies our required condition.
full solution