This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: BTRY/STSCI 4080 Homework # 10 Due Date: before noon on 12/6/10 Problem numbers beginning with P (e.g., P1) denote problems from the PROB LEMS section; those beginning with T (e.g., T1) denote problems taken from the THEORETICAL EXERCISES section. 1. Let a,b R , with a < b . Suppose g ( ) is a function that satisfies: g ( u ) > 0 and is continuous for u [ a,b ] max u [ a,b ] g ( u ) = M < (i.e., g ( u ) is bounded between [0 ,M ] for u [ a,b ]). Let X 1 ,...,X n be independent Uniform( a,b ) random variables. Let Y 1 ,...,Y n be indepen dent Uniform(0 ,M ) random variables. Suppose the X s and Y s are all mutually independent. Finally, define Z i = I { g ( X i ) > Y i } , and W n = M ( b a ) n n i =1 Z i . (a) Explain why Z 1 ...Z n are independent Bernoulli( p ) random variables, where p = P ( g ( X 1 ) > Y 1 ) . What is the probability distribution of n i =1 Z i ? (b) Show that P ( g ( X 1 ) > Y 1 ) = 1 M ( b a ) b a g ( x ) dx . (c) Show that E ( W n ) = I g and V ar ( W n ) = n 1 I g ( M ( b a )I g ), where I g = b a g ( u ) du. (d) Let g ( ) = ( ), where ( ) is the standard normal pdf. Consider the integral I g = 1 . 28 1 . 645 ( u ) du = 0 . 8497425 . Show that one can take M = (0) = 1 2 and compute V ar ( W n ) . How large must n be to ensure that SD ( W n ) < . 001?...
View
Full
Document
 Fall '10
 STRAWDERMAN

Click to edit the document details