In: Statistics and Probability
If T1 and T2 are independent exponential random variables, find the density function of R=T(2) - T(1).
This is for the difference of the order statistics not of the variables, i.e. we are not looking for T2 - T1. It is implied that they are both from the same distribution. I know that
fT(t) = λe-λt
fT(1)T(2)(t1,t2) = 2 fT(t1)fT(t2) = 2λ2 e-λt1 e-λt2 , 0 < t1 < t2 and I need to find fR(r).
From Mathematical Statistics and Data Analysis by Rice, 3rd Edition, Question 79 from Chapter 3. The solution in the back of the book: "Exponential (λ)". I am not sure how to get to that answer.
let be random samples from exponential distribution with
cdf
Using the formula for joint pdf of
if are the ordered statistics, the joint pdf is
Let and giving us 0<R<S
That is can be expressed in terms of S and R as
The Jacobian |J| is
the joint pdf of R,S is
The marginal pdf of R is
This is the pdf of exponential distribution with parameter
That means