Determine marginal densities and distributions from joint density

  • Thread starter Thread starter psie
  • Start date Start date
  • Tags Tags
    Probability theory
psie
Messages
315
Reaction score
40
Homework Statement
Let ##X## and ##Y## have joint density $$f(x,y)=\begin{cases} 1& \text{for }0\leq x\leq 2,\max(0,x-1)\leq y\leq\min(1,x) \\ 0 &\text{otherwise}.\end{cases}$$ Find the marginal density functions and the joint and marginal distribution functions.
Relevant Equations
The marginal distribution of ##X## given the joint density ##f_{X,Y}## is given by ##f_X(x)=\int_\mathbb{R} f_{X,Y}(x,y) \,dy## and similar for ##Y##.
This is the follow-up problem to my previous problem.

"Integrating out" the ##y##-variable and ##x##-variable separately, we see that ##f_Y(y)=2## and ##f_X(x)=\min(1,x)-\max(0,x-1)##. From my previous post, we see that ##X## is the sum of two independent ##U(0,1)##-distributed r.v.s. What is the distribution of ##Y## though? It looks to me that if ##0\leq x\leq 2##, i.e. if ##0\leq x\leq 1## or ##1<x\leq 2##, then ##0\leq y\leq x## or ##x-1\leq y\leq 1## respectively. So ##y## ranges from ##0## to ##1##, which doesn't make sense since then the pdf ##f_Y(y)=2## does not integrate to ##1##. However, currently I don't see the error.
 
Last edited:
Physics news on Phys.org
What makes you think that this
psie said:
$$f(x,y)=\begin{cases} 1& \text{for }0\leq x\leq 2,\max(0,x-1)\leq y\leq\min(1,x) \\ 0 &\text{otherwise}.\end{cases}$$
Is the density function of the independent RVs ##X## and ##Y## from your previous post?

If ##X## and ##Y## are uniform independent RVs then the density function is a constant on the product of the supports of each variable.
 
Additionally, even if the problem is not supposed to be the independent RVs from your previous post, the correct computation of ##f_Y(y)## has the result ##f_Y(y) = 1##, not 2.
 
Orodruin said:
What makes you think that this is the density function of the independent RVs ##X## and ##Y## from your previous post?
I am not saying that ##f## given here is the joint density of two ##U(0,1)##-distributed, independent RVs. What I'm claiming is that ##X## in this post is the sum of two independent RVs, both of which are ##U(0,1)##, since in my previous post, we found that the density of the sum of two independent RVs which are ##U(0,1)## is ##f_X## as specified here. Am I making sense?
Orodruin said:
Additionally, even if the problem is not supposed to be the independent RVs from your previous post, the correct computation of ##f_Y(y)## has the result ##f_Y(y) = 1##, not 2.
How did you obtain ##f_Y(y) = 1##? I just don't see it.
 
psie said:
How did you obtain ##f_Y(y) = 1##? I just don't see it.
Did you try drawing the support of the joint distribution function? If you do it should be pretty clear.
 
Also note that writing down the support region is significantly simpler if you change the order. In other words, start writing down the domain for ##y## and then write the restriction on ##x## given ##y##.
 
Orodruin said:
Did you try drawing the support of the joint distribution function? If you do it should be pretty clear.
Ok, I will try. What is wrong about calculating the marginal density by integrating out the ##x##-variable though? In other words, $$f_Y(y)=\int_\mathbb{R} f(x,y) \,dx=\int_0^2 1\, dx=2?$$
 
psie said:
Ok, I will try. What is wrong about calculating the marginal density by integrating out the ##x##-variable though? In other words, $$f_Y(y)=\int_\mathbb{R} f(x,y) \,dx=\int_0^2 1\, dx=2?$$
$$f_Y(y)=\int_\mathbb{R} f(x,y) \,dx \neq \int_0^2 1\, dx=2?$$

Draw the support region for ##f## and you will see it.
 
Back
Top