Variance of $XY(1-Y)$ in terms of the means and variances $X$ and $Y$
Consider two independent random variables X and Y. X has some distribution with mean $mu_X$ and variance $sigma^2_X$. Y has some distribution with mean $mu_Y$ and variance $sigma^2_Y$. I want to take the variance of the quantity $XY(1-Y)$. Is the correct answer:
$$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]?$$
I just did some kind of weird product rule but with variances and I'm not sure that's at all legit.
The reason I want to know the variance of this quantity is that it is a measure of the real noise in an experimental system I am working with. I am an experimental physicist and my math skills are weak at best. You might recognize the form $XY(1-Y)$ from the binomial distribution: $Np(1-p)$. My experiment does not follow a binomial distribution exactly, it follows a Poisson binomial distribution but with varying $N$ and $p$ for each shot of the experiment. So I need to know the variance of the quantity $XY(1-Y)$ when I know the variances and expectation values of $X$ and $Y$.
I'm trying to work this particular detail out to answer this question:
Repeated binomial processes with different probabilities and number of trials .
probability-theory variance
add a comment |
Consider two independent random variables X and Y. X has some distribution with mean $mu_X$ and variance $sigma^2_X$. Y has some distribution with mean $mu_Y$ and variance $sigma^2_Y$. I want to take the variance of the quantity $XY(1-Y)$. Is the correct answer:
$$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]?$$
I just did some kind of weird product rule but with variances and I'm not sure that's at all legit.
The reason I want to know the variance of this quantity is that it is a measure of the real noise in an experimental system I am working with. I am an experimental physicist and my math skills are weak at best. You might recognize the form $XY(1-Y)$ from the binomial distribution: $Np(1-p)$. My experiment does not follow a binomial distribution exactly, it follows a Poisson binomial distribution but with varying $N$ and $p$ for each shot of the experiment. So I need to know the variance of the quantity $XY(1-Y)$ when I know the variances and expectation values of $X$ and $Y$.
I'm trying to work this particular detail out to answer this question:
Repeated binomial processes with different probabilities and number of trials .
probability-theory variance
1
Comments are not for extended discussion; this conversation has been moved to chat.
– quid♦
Dec 24 '18 at 14:55
1
The formula $$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]$$ cannot be correct since $sigma^2_{XY(1-Y)}$, $sigma^2_X$ and $sigma^2_Y(1-sigma^2_Y)$ are numbers while $[Y(1-Y)]^2$ and $X^2$ are random variables.
– Did
Dec 25 '18 at 18:08
add a comment |
Consider two independent random variables X and Y. X has some distribution with mean $mu_X$ and variance $sigma^2_X$. Y has some distribution with mean $mu_Y$ and variance $sigma^2_Y$. I want to take the variance of the quantity $XY(1-Y)$. Is the correct answer:
$$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]?$$
I just did some kind of weird product rule but with variances and I'm not sure that's at all legit.
The reason I want to know the variance of this quantity is that it is a measure of the real noise in an experimental system I am working with. I am an experimental physicist and my math skills are weak at best. You might recognize the form $XY(1-Y)$ from the binomial distribution: $Np(1-p)$. My experiment does not follow a binomial distribution exactly, it follows a Poisson binomial distribution but with varying $N$ and $p$ for each shot of the experiment. So I need to know the variance of the quantity $XY(1-Y)$ when I know the variances and expectation values of $X$ and $Y$.
I'm trying to work this particular detail out to answer this question:
Repeated binomial processes with different probabilities and number of trials .
probability-theory variance
Consider two independent random variables X and Y. X has some distribution with mean $mu_X$ and variance $sigma^2_X$. Y has some distribution with mean $mu_Y$ and variance $sigma^2_Y$. I want to take the variance of the quantity $XY(1-Y)$. Is the correct answer:
$$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]?$$
I just did some kind of weird product rule but with variances and I'm not sure that's at all legit.
The reason I want to know the variance of this quantity is that it is a measure of the real noise in an experimental system I am working with. I am an experimental physicist and my math skills are weak at best. You might recognize the form $XY(1-Y)$ from the binomial distribution: $Np(1-p)$. My experiment does not follow a binomial distribution exactly, it follows a Poisson binomial distribution but with varying $N$ and $p$ for each shot of the experiment. So I need to know the variance of the quantity $XY(1-Y)$ when I know the variances and expectation values of $X$ and $Y$.
I'm trying to work this particular detail out to answer this question:
Repeated binomial processes with different probabilities and number of trials .
probability-theory variance
probability-theory variance
edited Dec 25 '18 at 18:09
Did
246k23221455
246k23221455
asked Dec 23 '18 at 5:42
SabrinaChoiceSabrinaChoice
246
246
1
Comments are not for extended discussion; this conversation has been moved to chat.
– quid♦
Dec 24 '18 at 14:55
1
The formula $$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]$$ cannot be correct since $sigma^2_{XY(1-Y)}$, $sigma^2_X$ and $sigma^2_Y(1-sigma^2_Y)$ are numbers while $[Y(1-Y)]^2$ and $X^2$ are random variables.
– Did
Dec 25 '18 at 18:08
add a comment |
1
Comments are not for extended discussion; this conversation has been moved to chat.
– quid♦
Dec 24 '18 at 14:55
1
The formula $$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]$$ cannot be correct since $sigma^2_{XY(1-Y)}$, $sigma^2_X$ and $sigma^2_Y(1-sigma^2_Y)$ are numbers while $[Y(1-Y)]^2$ and $X^2$ are random variables.
– Did
Dec 25 '18 at 18:08
1
1
Comments are not for extended discussion; this conversation has been moved to chat.
– quid♦
Dec 24 '18 at 14:55
Comments are not for extended discussion; this conversation has been moved to chat.
– quid♦
Dec 24 '18 at 14:55
1
1
The formula $$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]$$ cannot be correct since $sigma^2_{XY(1-Y)}$, $sigma^2_X$ and $sigma^2_Y(1-sigma^2_Y)$ are numbers while $[Y(1-Y)]^2$ and $X^2$ are random variables.
– Did
Dec 25 '18 at 18:08
The formula $$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]$$ cannot be correct since $sigma^2_{XY(1-Y)}$, $sigma^2_X$ and $sigma^2_Y(1-sigma^2_Y)$ are numbers while $[Y(1-Y)]^2$ and $X^2$ are random variables.
– Did
Dec 25 '18 at 18:08
add a comment |
1 Answer
1
active
oldest
votes
$begin{aligned}mathsf{Var}XYleft(1-Yright) & =mathsf{Covar}left(XYleft(1-Yright),XYleft(1-Yright)right)\
& =mathsf{Covar}left(XY-XY^{2},XY-XY^{2}right)\
& =mathsf{Var}left(XYright)-2mathsf{Covar}left(XY^{2},XYright)+mathsf{Var}left(XY^{2}right)\
& =left[mathbb{E}X^{2}Y^{2}-left(mathbb{E}XYright)^{2}right]-2left[mathbb{E}X^{2}Y^{3}-mathbb{E}XY^{2}mathbb{E}XYright]+left[mathbb{E}X^{2}Y^{4}-left(mathbb{E}XY^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-left(mathbb{E}Xmathbb{E}Yright)^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mathbb{E}Xmathbb{E}Y^{2}mathbb{E}Xmathbb{E}Yright]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-left(mathbb{E}Xmathbb{E}Y^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-mu_{X}^{2}mu_{Y}^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}mathbb{E}Y^{2}right]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-mu_{X}^{2}left(mathbb{E}Y^{2}right)^{2}right]\
& =left[left(sigma_{X}^{2}+mu_{X}^{2}right)left(sigma_{Y}^{2}+mu_{Y}^{2}right)-mu_{X}^{2}mu_{Y}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]\
& =left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]
end{aligned}
$
Especially the $5$-th equality rests on independence of $X$ and $Y$.
We cannot get any further since for the expressions $mathbb EY^3$ and $mathbb EY^4$ more information is needed concerning the distribution of $Y$.
edit:
Special case: $Ysimmathsf{Normal}left(mu_{Y},sigma_{Y}^{2}right)$.
Then $Y=mu_{Y}+sigma_{Y}U$ where $Usimmathsf{Normal}left(0,1right)$
and based on the knowledge that $mathbb{E}U^{3}=0$ and $mathbb{E}U^{4}=3$
we find:
- $mathbb{E}Y^{3}=mathbb{E}left(mu_{Y}^{3}+3mu_{Y}^{2}sigma_{Y}U+3mu_{Y}sigma_{Y}^{2}U^{2}+sigma_{Y}^{3}U^{3}right)=mu_{Y}^{3}+3mu_{Y}sigma_{Y}^{2}$
- $mathbb{E}Y^{4}=mathbb{E}left(mu_{Y}^{4}+4mu_{Y}^{3}sigma_{Y}U+6mu_{Y}^{2}sigma_{Y}^{2}U^{2}+4mu_{Y}sigma_{Y}^{3}U^{3}+sigma_{Y}^{4}U^{4}right)=mu_{Y}^{4}+6mu_{Y}^{2}sigma_{Y}^{2}+3sigma_{Y}^{4}$
Also have a look here for that.
Substituting we find the following expression for $mathsf{Var}XYleft(1-Yright)$:
$$left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[sigma_{X}^{2}mu_{Y}^{3}+3sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}+2sigma_{Y}^{2}mu_{X}^{2}mu_{Y}right]+left[sigma_{X}^{2}mu_{Y}^{4}+6sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}^{2}+3sigma_{X}^{2}sigma_{Y}^{4}+4sigma_{Y}^{2}mu_{X}^{2}mu_{Y}^{2}+2sigma_{Y}^{4}mu_{X}^{2}right]$$
I hope I did not make any mistakes (check me on it).
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
1
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
1
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
1
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
|
show 2 more comments
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "69"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3050093%2fvariance-of-xy1-y-in-terms-of-the-means-and-variances-x-and-y%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begin{aligned}mathsf{Var}XYleft(1-Yright) & =mathsf{Covar}left(XYleft(1-Yright),XYleft(1-Yright)right)\
& =mathsf{Covar}left(XY-XY^{2},XY-XY^{2}right)\
& =mathsf{Var}left(XYright)-2mathsf{Covar}left(XY^{2},XYright)+mathsf{Var}left(XY^{2}right)\
& =left[mathbb{E}X^{2}Y^{2}-left(mathbb{E}XYright)^{2}right]-2left[mathbb{E}X^{2}Y^{3}-mathbb{E}XY^{2}mathbb{E}XYright]+left[mathbb{E}X^{2}Y^{4}-left(mathbb{E}XY^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-left(mathbb{E}Xmathbb{E}Yright)^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mathbb{E}Xmathbb{E}Y^{2}mathbb{E}Xmathbb{E}Yright]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-left(mathbb{E}Xmathbb{E}Y^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-mu_{X}^{2}mu_{Y}^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}mathbb{E}Y^{2}right]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-mu_{X}^{2}left(mathbb{E}Y^{2}right)^{2}right]\
& =left[left(sigma_{X}^{2}+mu_{X}^{2}right)left(sigma_{Y}^{2}+mu_{Y}^{2}right)-mu_{X}^{2}mu_{Y}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]\
& =left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]
end{aligned}
$
Especially the $5$-th equality rests on independence of $X$ and $Y$.
We cannot get any further since for the expressions $mathbb EY^3$ and $mathbb EY^4$ more information is needed concerning the distribution of $Y$.
edit:
Special case: $Ysimmathsf{Normal}left(mu_{Y},sigma_{Y}^{2}right)$.
Then $Y=mu_{Y}+sigma_{Y}U$ where $Usimmathsf{Normal}left(0,1right)$
and based on the knowledge that $mathbb{E}U^{3}=0$ and $mathbb{E}U^{4}=3$
we find:
- $mathbb{E}Y^{3}=mathbb{E}left(mu_{Y}^{3}+3mu_{Y}^{2}sigma_{Y}U+3mu_{Y}sigma_{Y}^{2}U^{2}+sigma_{Y}^{3}U^{3}right)=mu_{Y}^{3}+3mu_{Y}sigma_{Y}^{2}$
- $mathbb{E}Y^{4}=mathbb{E}left(mu_{Y}^{4}+4mu_{Y}^{3}sigma_{Y}U+6mu_{Y}^{2}sigma_{Y}^{2}U^{2}+4mu_{Y}sigma_{Y}^{3}U^{3}+sigma_{Y}^{4}U^{4}right)=mu_{Y}^{4}+6mu_{Y}^{2}sigma_{Y}^{2}+3sigma_{Y}^{4}$
Also have a look here for that.
Substituting we find the following expression for $mathsf{Var}XYleft(1-Yright)$:
$$left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[sigma_{X}^{2}mu_{Y}^{3}+3sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}+2sigma_{Y}^{2}mu_{X}^{2}mu_{Y}right]+left[sigma_{X}^{2}mu_{Y}^{4}+6sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}^{2}+3sigma_{X}^{2}sigma_{Y}^{4}+4sigma_{Y}^{2}mu_{X}^{2}mu_{Y}^{2}+2sigma_{Y}^{4}mu_{X}^{2}right]$$
I hope I did not make any mistakes (check me on it).
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
1
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
1
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
1
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
|
show 2 more comments
$begin{aligned}mathsf{Var}XYleft(1-Yright) & =mathsf{Covar}left(XYleft(1-Yright),XYleft(1-Yright)right)\
& =mathsf{Covar}left(XY-XY^{2},XY-XY^{2}right)\
& =mathsf{Var}left(XYright)-2mathsf{Covar}left(XY^{2},XYright)+mathsf{Var}left(XY^{2}right)\
& =left[mathbb{E}X^{2}Y^{2}-left(mathbb{E}XYright)^{2}right]-2left[mathbb{E}X^{2}Y^{3}-mathbb{E}XY^{2}mathbb{E}XYright]+left[mathbb{E}X^{2}Y^{4}-left(mathbb{E}XY^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-left(mathbb{E}Xmathbb{E}Yright)^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mathbb{E}Xmathbb{E}Y^{2}mathbb{E}Xmathbb{E}Yright]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-left(mathbb{E}Xmathbb{E}Y^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-mu_{X}^{2}mu_{Y}^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}mathbb{E}Y^{2}right]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-mu_{X}^{2}left(mathbb{E}Y^{2}right)^{2}right]\
& =left[left(sigma_{X}^{2}+mu_{X}^{2}right)left(sigma_{Y}^{2}+mu_{Y}^{2}right)-mu_{X}^{2}mu_{Y}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]\
& =left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]
end{aligned}
$
Especially the $5$-th equality rests on independence of $X$ and $Y$.
We cannot get any further since for the expressions $mathbb EY^3$ and $mathbb EY^4$ more information is needed concerning the distribution of $Y$.
edit:
Special case: $Ysimmathsf{Normal}left(mu_{Y},sigma_{Y}^{2}right)$.
Then $Y=mu_{Y}+sigma_{Y}U$ where $Usimmathsf{Normal}left(0,1right)$
and based on the knowledge that $mathbb{E}U^{3}=0$ and $mathbb{E}U^{4}=3$
we find:
- $mathbb{E}Y^{3}=mathbb{E}left(mu_{Y}^{3}+3mu_{Y}^{2}sigma_{Y}U+3mu_{Y}sigma_{Y}^{2}U^{2}+sigma_{Y}^{3}U^{3}right)=mu_{Y}^{3}+3mu_{Y}sigma_{Y}^{2}$
- $mathbb{E}Y^{4}=mathbb{E}left(mu_{Y}^{4}+4mu_{Y}^{3}sigma_{Y}U+6mu_{Y}^{2}sigma_{Y}^{2}U^{2}+4mu_{Y}sigma_{Y}^{3}U^{3}+sigma_{Y}^{4}U^{4}right)=mu_{Y}^{4}+6mu_{Y}^{2}sigma_{Y}^{2}+3sigma_{Y}^{4}$
Also have a look here for that.
Substituting we find the following expression for $mathsf{Var}XYleft(1-Yright)$:
$$left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[sigma_{X}^{2}mu_{Y}^{3}+3sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}+2sigma_{Y}^{2}mu_{X}^{2}mu_{Y}right]+left[sigma_{X}^{2}mu_{Y}^{4}+6sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}^{2}+3sigma_{X}^{2}sigma_{Y}^{4}+4sigma_{Y}^{2}mu_{X}^{2}mu_{Y}^{2}+2sigma_{Y}^{4}mu_{X}^{2}right]$$
I hope I did not make any mistakes (check me on it).
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
1
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
1
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
1
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
|
show 2 more comments
$begin{aligned}mathsf{Var}XYleft(1-Yright) & =mathsf{Covar}left(XYleft(1-Yright),XYleft(1-Yright)right)\
& =mathsf{Covar}left(XY-XY^{2},XY-XY^{2}right)\
& =mathsf{Var}left(XYright)-2mathsf{Covar}left(XY^{2},XYright)+mathsf{Var}left(XY^{2}right)\
& =left[mathbb{E}X^{2}Y^{2}-left(mathbb{E}XYright)^{2}right]-2left[mathbb{E}X^{2}Y^{3}-mathbb{E}XY^{2}mathbb{E}XYright]+left[mathbb{E}X^{2}Y^{4}-left(mathbb{E}XY^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-left(mathbb{E}Xmathbb{E}Yright)^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mathbb{E}Xmathbb{E}Y^{2}mathbb{E}Xmathbb{E}Yright]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-left(mathbb{E}Xmathbb{E}Y^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-mu_{X}^{2}mu_{Y}^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}mathbb{E}Y^{2}right]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-mu_{X}^{2}left(mathbb{E}Y^{2}right)^{2}right]\
& =left[left(sigma_{X}^{2}+mu_{X}^{2}right)left(sigma_{Y}^{2}+mu_{Y}^{2}right)-mu_{X}^{2}mu_{Y}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]\
& =left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]
end{aligned}
$
Especially the $5$-th equality rests on independence of $X$ and $Y$.
We cannot get any further since for the expressions $mathbb EY^3$ and $mathbb EY^4$ more information is needed concerning the distribution of $Y$.
edit:
Special case: $Ysimmathsf{Normal}left(mu_{Y},sigma_{Y}^{2}right)$.
Then $Y=mu_{Y}+sigma_{Y}U$ where $Usimmathsf{Normal}left(0,1right)$
and based on the knowledge that $mathbb{E}U^{3}=0$ and $mathbb{E}U^{4}=3$
we find:
- $mathbb{E}Y^{3}=mathbb{E}left(mu_{Y}^{3}+3mu_{Y}^{2}sigma_{Y}U+3mu_{Y}sigma_{Y}^{2}U^{2}+sigma_{Y}^{3}U^{3}right)=mu_{Y}^{3}+3mu_{Y}sigma_{Y}^{2}$
- $mathbb{E}Y^{4}=mathbb{E}left(mu_{Y}^{4}+4mu_{Y}^{3}sigma_{Y}U+6mu_{Y}^{2}sigma_{Y}^{2}U^{2}+4mu_{Y}sigma_{Y}^{3}U^{3}+sigma_{Y}^{4}U^{4}right)=mu_{Y}^{4}+6mu_{Y}^{2}sigma_{Y}^{2}+3sigma_{Y}^{4}$
Also have a look here for that.
Substituting we find the following expression for $mathsf{Var}XYleft(1-Yright)$:
$$left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[sigma_{X}^{2}mu_{Y}^{3}+3sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}+2sigma_{Y}^{2}mu_{X}^{2}mu_{Y}right]+left[sigma_{X}^{2}mu_{Y}^{4}+6sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}^{2}+3sigma_{X}^{2}sigma_{Y}^{4}+4sigma_{Y}^{2}mu_{X}^{2}mu_{Y}^{2}+2sigma_{Y}^{4}mu_{X}^{2}right]$$
I hope I did not make any mistakes (check me on it).
$begin{aligned}mathsf{Var}XYleft(1-Yright) & =mathsf{Covar}left(XYleft(1-Yright),XYleft(1-Yright)right)\
& =mathsf{Covar}left(XY-XY^{2},XY-XY^{2}right)\
& =mathsf{Var}left(XYright)-2mathsf{Covar}left(XY^{2},XYright)+mathsf{Var}left(XY^{2}right)\
& =left[mathbb{E}X^{2}Y^{2}-left(mathbb{E}XYright)^{2}right]-2left[mathbb{E}X^{2}Y^{3}-mathbb{E}XY^{2}mathbb{E}XYright]+left[mathbb{E}X^{2}Y^{4}-left(mathbb{E}XY^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-left(mathbb{E}Xmathbb{E}Yright)^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mathbb{E}Xmathbb{E}Y^{2}mathbb{E}Xmathbb{E}Yright]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-left(mathbb{E}Xmathbb{E}Y^{2}right)^{2}right]\
& =left[mathbb{E}X^{2}mathbb{E}Y^{2}-mu_{X}^{2}mu_{Y}^{2}right]-2left[mathbb{E}X^{2}mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}mathbb{E}Y^{2}right]+left[mathbb{E}X^{2}mathbb{E}Y^{4}-mu_{X}^{2}left(mathbb{E}Y^{2}right)^{2}right]\
& =left[left(sigma_{X}^{2}+mu_{X}^{2}right)left(sigma_{Y}^{2}+mu_{Y}^{2}right)-mu_{X}^{2}mu_{Y}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]\
& =left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{3}-mu_{X}^{2}mu_{Y}left(sigma_{Y}^{2}+mu_{Y}^{2}right)right]+left[left(sigma_{X}^{2}+mu_{X}^{2}right)mathbb{E}Y^{4}-mu_{X}^{2}left(sigma_{Y}^{2}+mu_{Y}^{2}right)^{2}right]
end{aligned}
$
Especially the $5$-th equality rests on independence of $X$ and $Y$.
We cannot get any further since for the expressions $mathbb EY^3$ and $mathbb EY^4$ more information is needed concerning the distribution of $Y$.
edit:
Special case: $Ysimmathsf{Normal}left(mu_{Y},sigma_{Y}^{2}right)$.
Then $Y=mu_{Y}+sigma_{Y}U$ where $Usimmathsf{Normal}left(0,1right)$
and based on the knowledge that $mathbb{E}U^{3}=0$ and $mathbb{E}U^{4}=3$
we find:
- $mathbb{E}Y^{3}=mathbb{E}left(mu_{Y}^{3}+3mu_{Y}^{2}sigma_{Y}U+3mu_{Y}sigma_{Y}^{2}U^{2}+sigma_{Y}^{3}U^{3}right)=mu_{Y}^{3}+3mu_{Y}sigma_{Y}^{2}$
- $mathbb{E}Y^{4}=mathbb{E}left(mu_{Y}^{4}+4mu_{Y}^{3}sigma_{Y}U+6mu_{Y}^{2}sigma_{Y}^{2}U^{2}+4mu_{Y}sigma_{Y}^{3}U^{3}+sigma_{Y}^{4}U^{4}right)=mu_{Y}^{4}+6mu_{Y}^{2}sigma_{Y}^{2}+3sigma_{Y}^{4}$
Also have a look here for that.
Substituting we find the following expression for $mathsf{Var}XYleft(1-Yright)$:
$$left[sigma_{X}^{2}sigma_{Y}^{2}+sigma_{X}^{2}mu_{Y}^{2}+sigma_{Y}^{2}mu_{X}^{2}right]-2left[sigma_{X}^{2}mu_{Y}^{3}+3sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}+2sigma_{Y}^{2}mu_{X}^{2}mu_{Y}right]+left[sigma_{X}^{2}mu_{Y}^{4}+6sigma_{X}^{2}sigma_{Y}^{2}mu_{Y}^{2}+3sigma_{X}^{2}sigma_{Y}^{4}+4sigma_{Y}^{2}mu_{X}^{2}mu_{Y}^{2}+2sigma_{Y}^{4}mu_{X}^{2}right]$$
I hope I did not make any mistakes (check me on it).
edited Jan 4 at 8:06
answered Dec 24 '18 at 13:11
drhabdrhab
98.4k544129
98.4k544129
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
1
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
1
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
1
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
|
show 2 more comments
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
1
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
1
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
1
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
This is wonderful, thank you! I wouldn't have thought of taking the covariance first. I'm not actually sure about the exact distributions of X and Y (they will be empirically determined). They'll probably end up Gaussian. Once I know them, then I would just have to do the integrals. Thanks so much, this is really helpful.
– SabrinaChoice
Dec 24 '18 at 14:47
1
1
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
@SabrinaChoice: In general, you can try to look at the moments of any sufficiently nice distribution if they exist, and each moment is a separate 'degree of freedom'. In fact, if all finite moments exist, then they uniquely characterize the distribution (if it has a continuous CDF). So we can also compare distributions by comparing their moments.
– user21820
Dec 24 '18 at 15:02
1
1
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
@SabrinaChoice I let you know that I found a mistake (and repaired). $mathbb EX^2mathbb EX^2$ was accidently replaced by $sigma_X^2+mu_X^2$, but had to be replaced by $(sigma_X^2+mu_X^2)(sigma_Y^2+mu_Y^2)$ of course.
– drhab
Dec 25 '18 at 16:44
1
1
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
In my former comment $mathbb EX^2mathbb EX^2$ should be $mathbb EX^2mathbb EY^2$. Also I edited for the special case that $Y$ takes normal distribution.
– drhab
Dec 26 '18 at 8:15
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
@drhab I finally had the chance to go through this carefully and check for errors. I couldn't find any. Thank you very much, I really appreciate it. My math skills are weak, but I feel like this particular example will make me much better equipped to solve a problem like this on my own next time I encounter one.
– SabrinaChoice
Jan 4 at 6:52
|
show 2 more comments
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3050093%2fvariance-of-xy1-y-in-terms-of-the-means-and-variances-x-and-y%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
1
Comments are not for extended discussion; this conversation has been moved to chat.
– quid♦
Dec 24 '18 at 14:55
1
The formula $$sigma^2_{XY(1-Y)} = sigma^2_X[Y(1-Y)]^2+X^2[sigma^2_Y(1-sigma^2_Y)]$$ cannot be correct since $sigma^2_{XY(1-Y)}$, $sigma^2_X$ and $sigma^2_Y(1-sigma^2_Y)$ are numbers while $[Y(1-Y)]^2$ and $X^2$ are random variables.
– Did
Dec 25 '18 at 18:08