Register to reply

Intuitive reason absolute values are used for transformations in statistics?

Share this thread:
phiiota
#1
Nov30-12, 05:52 PM
P: 29
1. The problem statement, all variables and given/known data



2. Relevant equations



3. The attempt at a solution

this isn't really homework, but I was just wondering if someone could offer an intuitive reason as to why when random variables are transformed, we use absolute values of derivative of those functions, as opposed to the functions themselves?
Phys.Org News Partner Science news on Phys.org
Bees able to spot which flowers offer best rewards before landing
Classic Lewis Carroll character inspires new ecological model
When cooperation counts: Researchers find sperm benefit from grouping together in mice
Mark44
#2
Nov30-12, 08:18 PM
Mentor
P: 21,215
Quote Quote by phiiota View Post
1. The problem statement, all variables and given/known data



2. Relevant equations



3. The attempt at a solution

this isn't really homework, but I was just wondering if someone could offer an intuitive reason as to why when random variables are transformed, we use absolute values of derivative of those functions, as opposed to the functions themselves?
Can you give an example where this is happening?
phiiota
#3
Nov30-12, 09:38 PM
P: 29
well, if say X distributed as f(X), and say i have Y=g(X), and I want to know the distribution of Y, in a simple case I can say that Y~f(g-1(y))|d/dy g-1(y)|.

I don't know why we'd always want the absolute value, as opposed to the derivative in general.

phiiota
#4
Nov30-12, 09:45 PM
P: 29
Intuitive reason absolute values are used for transformations in statistics?

or did you want a specific case? say X~U(0,1), and I want to know say Y=X-1.
then I can say g-1(y)=y-1., and d/dy g-1(y)=-y2, so then my distribution of Y would be

f(g-1(y))|d/dy g-1(y)|=y-2.

I get that if we didn't take the absolute value, then this function would be negative... But aside from that, I'm not seeing an intuitive idea as to why we'd always take the absolute value.
Ray Vickson
#5
Nov30-12, 10:03 PM
Sci Advisor
HW Helper
Thanks
P: 4,959
Quote Quote by phiiota View Post
well, if say X distributed as f(X), and say i have Y=g(X), and I want to know the distribution of Y, in a simple case I can say that Y~f(g-1(y))|d/dy g-1(y)|.

I don't know why we'd always want the absolute value, as opposed to the derivative in general.
[tex]P\{y < Y < y + dy\} = P\{y < g(X) < y + dy\}.[/tex] If g is an increasing function we have
[tex] P\{y < g(X) < y+dy\} = P\{g^{-1}(y) < X < g^{-1}(y+dy) \} =
P\{ g^{-1}(y) < X < g^{-1}(y) + {g^{-1}}^{\prime} (y) dy\} \doteq f[g^{-1}(y)] {g^{-1}}^{\prime}(y) \, dy.[/tex]
If g is a decreasing function we have
[tex] P\{y < g(X) < y+dy\} = P \{ g^{-1}(y+dy) < X < g^{-1}(y) \}
= P\{ g^{-1}(y) - |{g^{-1}}^{\prime}(y)| \, dy < X < g^{-1}(y) \}
\doteq f[g^{-1}(y)]\, | {g^{-1}}^{\prime}(y) | \, dy.[/tex] The point is that the *length* of the x-interval corresponding to dy is [itex] | {g^{-1}}^{\prime}(y) | \, dy [/itex], and you need the absolute value so that the length cannot be negative.
phiiota
#6
Nov30-12, 10:09 PM
P: 29
Okay, that makes sense. Thank you.


Register to reply

Related Discussions
B-E,F-D Statistics and the Lorentz Transformations Special & General Relativity 7
Causal structure of spacetime: intuitive reason for non-compactness? Special & General Relativity 4
Intuitive reason for no gravity in a hollow sphere Classical Physics 29
Help On Absolute Values Introductory Physics Homework 3
Absolute values Introductory Physics Homework 12