Results 1 to 5 of 5

Math Help - Method of Distribution Functions problem

  1. #1
    Newbie
    Joined
    Apr 2013
    From
    Canada
    Posts
    7

    Method of Distribution Functions problem

    I need help with this problem:

    "The joint distribution of amount of pollutant emmited from a smokestack without cleaning device(Y_1) and a similar smokestack with a cleaning device (Y_2) is

    f(y_1,y_2)= 1, 0<=(less or equal to)y_1<=2, 0<=y_2<=1, 2y_2<=y_1
    0, elsewhere

    U = Y_1 - Y_2

    Find the probability density function for U"

    So I need to find F_U(u)=P(U<=u)=P(Y_1-Y_2<=u)=P(Y_1<=u+Y_2)

    I drew the attachment, but I don't know how to proceed form there. I don't understand how to choose the region U<=u or where u splits.
    I know for u<0, F_U(u)=0 since Y_1-Y_2<0 has no solution and for u>2, F_U(u)=1
    Attached Thumbnails Attached Thumbnails Method of Distribution Functions problem-problem.png  
    Last edited by Migno; April 21st 2013 at 02:38 PM.
    Follow Math Help Forum on Facebook and Google+

  2. #2
    MHF Contributor
    Joined
    Sep 2012
    From
    Australia
    Posts
    3,655
    Thanks
    604

    Re: Method of Distribution Functions problem

    Hey Migno.

    Hint: Using your information about the joint distribution of Y1 and Y2, where is Y1 - Y2 defined? (This will give you the region for the random variable and will help you get its distribution).
    Follow Math Help Forum on Facebook and Google+

  3. #3
    Newbie
    Joined
    Apr 2013
    From
    Canada
    Posts
    7

    Re: Method of Distribution Functions problem

    Well, 0<Y1-Y2<2 so 0<u<2
    I guess that's the region for U and if I let Y2=0 I get the intercept u=1, so I guess that's where it splits.
    So for 0<u<1 the region is above the line Y1=u+Y2 and for 1<u<2 the region is below that line. Is that right?
    Follow Math Help Forum on Facebook and Google+

  4. #4
    MHF Contributor
    Joined
    Sep 2012
    From
    Australia
    Posts
    3,655
    Thanks
    604

    Re: Method of Distribution Functions problem

    Now you can integrate over that region with respect to a varying value of u (with your limits) to get a CDF and differentiate to get a PDF in terms of u.

    Can you write an integral in terms of u given the limits of joint distribution that satisfy Y1 - Y2 being in the set?

    (Hint: Think about each value of u as a slice for each valid value of Y1 - Y2 in the integration region and the probability for a particular u sums up all the probabilities in that slice. To sum something up, you integrate the probability for that slice and that becomes your PDF for P(U = u)).
    Follow Math Help Forum on Facebook and Google+

  5. #5
    Newbie
    Joined
    Apr 2013
    From
    Canada
    Posts
    7

    Re: Method of Distribution Functions problem

    Ok, so the drawing was bad and it was throwing me off so I attached a better one. So the area Y_1-Y_2<u is always above the line Y_1-Y_2=u and with Y_2=0 we have that u=Y_1. After the line sort of passes the point (2,1) though, the integration limits will not be the same and at that point u=1, so that's why it has to split there.

    a) So for 0<u<1 we have the limits 0<y_2<u and 2y_2<y_1<u+y_2 for 1dy_1dy_2. Integrating part is easy so F_U(u)=(u^2)/2

    b) and for 1<u<2 the upper area gets complicated to integrate so you take the complement (A in the attachment) the limits are y_2+u<y_1<2 and 0<y_2<2-u again for 1dy_1dy_2. So F_U(u)=1-((2-u)^2)/2, the one minus is from the fact that it is the compliment.

    You differentiate those and get the pdf for u.

    Thanks chiro for helping me out. I thought I would share this since it took me forever to figure it out.
    Attached Thumbnails Attached Thumbnails Method of Distribution Functions problem-problem.png  
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Replies: 2
    Last Post: March 23rd 2013, 04:35 PM
  2. Replies: 1
    Last Post: March 21st 2013, 02:59 AM
  3. Need help with binomial distribution using the nCr method.
    Posted in the Advanced Statistics Forum
    Replies: 2
    Last Post: October 1st 2010, 11:51 AM
  4. Bi-nomial distribution - shorter method?
    Posted in the Statistics Forum
    Replies: 2
    Last Post: April 26th 2010, 04:31 PM
  5. Replies: 3
    Last Post: November 3rd 2007, 01:43 PM

Search Tags


/mathhelpforum @mathhelpforum