This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Pstat160A Winter 2009 Solution Homework #3 Problem 1. Form discussion session #2, problem 2 we have: G X ( s ) = G X,Y ( s, 1) so G X ( s ) = exp { λ 1 ( s 1) } (1) G Y ( s ) = G X,Y (1 ,s ) so G Y ( s ) = exp { λ 2 ( s 1) } (2) G X + Y ( s ) = G X,Y ( s,s ) so G X + Y ( s ) exp { ( λ 1 + λ 2 )( s 1) } (3) From (1) we recognize that X ∼ P ( λ 1 ), from (2) that Y ∼ P ( λ 2 ), and from (3) that X + Y ∼ P ( λ 1 + λ 2 ). However, X and Y are NOT independent (can’t split the generating function as product) unless γ = 0. If we compare to property as seen in class: sum of two independent Poisson is Poisson (with same parameters as here). The independence is not a necessary condition, or put differently, the converse is not true: If we know only the marginals are Poisson and the sum is Poisson (even with the “right” parameter), it doesn’t mean that the joint distribution is independent Poisson....
View
Full Document
 Spring '09
 Probability theory, Inverse transform sampling, inverse transform method, Pstat160A Solution Homework, Form discussion session, right distribution Problem

Click to edit the document details