Finding the P(2x1>x2) for a bivariate normal distribution

Click For Summary

Homework Help Overview

The discussion revolves around finding the probability P(2*x1>x2) for a bivariate normal distribution characterized by specific means and variances. The participants are exploring the properties of the distribution and the implications of its parameters.

Discussion Character

  • Exploratory, Mathematical reasoning, Assumption checking

Approaches and Questions Raised

  • Participants discuss the use of the cumulative distribution function (cdf) for the bivariate normal distribution and consider integration techniques to find the desired probability. There are attempts to express the cdf in terms of the variables involved and to manipulate the equation for further analysis.

Discussion Status

Some participants express uncertainty about their progress and whether they are missing key insights. Others share their attempts at deriving expressions for z and discuss the transformations involved. A participant indicates they have resolved their confusion, suggesting that productive dialogue has occurred.

Contextual Notes

There is mention of difficulties in performing the necessary integrations and concerns about the accuracy of the derived expressions. The original poster and others are working within the constraints of the problem as posed, with specific values for means and variances provided.

ahuds001
Messages
3
Reaction score
0

Homework Statement



Given a bivariate normal distribution with E(x1)=4 and E(x2) = 6 and Var(X) = [3 2.5]
[2.5 7]
Find P(2*x1>x2)

Homework Equations



The cdf of this bivariate normal distribution is given by:

f(x1,x2)=1/(2*pi*var(x1)*var(x2)*sqrt(1-rho^2)) * e^(-0.5*(z/(2*1-rho^2)))

where var(x1) = 3, var(x2) = 7, E(x1)=4 and E(x2) = 6, and rho = 2.5/(sqrt(3)*sqrt(7))

and z = ((x1-E(x1))^2)/var(x1) + ((x2-E(x2))^2)/var(x2) - (2*rho*(x1-E(x1))*(x2-E(x2)))/sqrt(var(x1)*var(x2))

The Attempt at a Solution



The approximate cdf of this bivariate normal distribution is given (in terms of x and y) by:

.041440417*%e^(-0.71186*((((x-4)^2)/3)+(((y-6)^2)/7)-(1.09109*(x-4)*(y-6))/4.582576))

Taking the integral from (-infinity to 2*x) dy and then from (-infinity to + infinity) dx should do the trick, but I have been unable to do so even using approximations.
 
Physics news on Phys.org
ahuds001 said:

Homework Statement



Given a bivariate normal distribution with E(x1)=4 and E(x2) = 6 and Var(X) = [3 2.5]
[2.5 7]
Find P(2*x1>x2)

Homework Equations



The cdf of this bivariate normal distribution is given by:

f(x1,x2)=1/(2*pi*var(x1)*var(x2)*sqrt(1-rho^2)) * e^(-0.5*(z/(2*1-rho^2)))

where var(x1) = 3, var(x2) = 7, E(x1)=4 and E(x2) = 6, and rho = 2.5/(sqrt(3)*sqrt(7))

and z = ((x1-E(x1))^2)/var(x1) + ((x2-E(x2))^2)/var(x2) - (2*rho*(x1-E(x1))*(x2-E(x2)))/sqrt(var(x1)*var(x2))

The Attempt at a Solution



The approximate cdf of this bivariate normal distribution is given (in terms of x and y) by:

.041440417*%e^(-0.71186*((((x-4)^2)/3)+(((y-6)^2)/7)-(1.09109*(x-4)*(y-6))/4.582576))

Taking the integral from (-infinity to 2*x) dy and then from (-infinity to + infinity) dx should do the trick, but I have been unable to do so even using approximations.

Can you figure out the distribution of the single random variable Y = 2*X1 - X2?

RGV
 
I can't, not sure if I am missing something or just being thick.
Working only on z=((x1-4)^2)/3 + ((x2-6)^2)/7 - 5*(x1-4)(x2-6)/21

Distributing I get z=(7x1^2+3x2^2-26x1-16x2-5x1x2+100)/21

My best attempt has z = (2y^2-4y-x1^2+x2^2+3x1x2-20x2-18x1+100)/21
 
Figured it out, thanks for the idea, it was extremely helpful :-)
 

Similar threads

  • · Replies 8 ·
Replies
8
Views
3K
Replies
14
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
Replies
5
Views
2K
Replies
8
Views
2K
  • · Replies 9 ·
Replies
9
Views
4K
Replies
8
Views
4K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K