Bound on third moment when the variable is bounded and has finite expectation
If I have a random variable $X$ such that $|X| leq 1$ and $mathbb{E}(X) = c < infty$, I want to find the tightest upper bound for $mathbb{E}(X^3)$. What would be the most optimal bound to use?
probability-theory
add a comment |
If I have a random variable $X$ such that $|X| leq 1$ and $mathbb{E}(X) = c < infty$, I want to find the tightest upper bound for $mathbb{E}(X^3)$. What would be the most optimal bound to use?
probability-theory
Let $P(X=1)=c$ and $P(X=0)=1-c$. Then $E(X)=E(X^3)=c$. I suspect this is optimal. In general you need to know the distribution of $X$.
– herb steinberg
Nov 30 '18 at 21:15
add a comment |
If I have a random variable $X$ such that $|X| leq 1$ and $mathbb{E}(X) = c < infty$, I want to find the tightest upper bound for $mathbb{E}(X^3)$. What would be the most optimal bound to use?
probability-theory
If I have a random variable $X$ such that $|X| leq 1$ and $mathbb{E}(X) = c < infty$, I want to find the tightest upper bound for $mathbb{E}(X^3)$. What would be the most optimal bound to use?
probability-theory
probability-theory
asked Nov 30 '18 at 19:23
Hanae SakurabaHanae Sakuraba
82
82
Let $P(X=1)=c$ and $P(X=0)=1-c$. Then $E(X)=E(X^3)=c$. I suspect this is optimal. In general you need to know the distribution of $X$.
– herb steinberg
Nov 30 '18 at 21:15
add a comment |
Let $P(X=1)=c$ and $P(X=0)=1-c$. Then $E(X)=E(X^3)=c$. I suspect this is optimal. In general you need to know the distribution of $X$.
– herb steinberg
Nov 30 '18 at 21:15
Let $P(X=1)=c$ and $P(X=0)=1-c$. Then $E(X)=E(X^3)=c$. I suspect this is optimal. In general you need to know the distribution of $X$.
– herb steinberg
Nov 30 '18 at 21:15
Let $P(X=1)=c$ and $P(X=0)=1-c$. Then $E(X)=E(X^3)=c$. I suspect this is optimal. In general you need to know the distribution of $X$.
– herb steinberg
Nov 30 '18 at 21:15
add a comment |
1 Answer
1
active
oldest
votes
Let $f(x)=x^3$ and let $Ssubset mathbb R^2$ be the convex hull of the points ${(x,f(x)): xin[-1,1]}$. The set $S$ is the set of all possible $(EX,EX^3)$ values for random variables $X$ with $P(|X|le1)=1.$ The intersection of $S$ with the vertical line cut out by $x=c$ is the interval from $(c,L(c))$ to $(c,U(c))$, where $L$ and $U$ are the lower convex and upper concave envelopes of $f$ on $[-1,1]$. Then the desired upper bound is $U(c)$.
More explicitly: Draw a line through $(1,1)$, tangent at $(-1/2, -1/8)$ to the graph of $f$. The graph of $U$ follows that of $f$ for $xin[-1,-1/2]$ and that of the tangent line for $xin[-1/2,1]$. The formula for $U$ is $$U(x)=begin{cases}-x^3&xle -1/2\ -frac 1 8 +frac 3 4 (x+1/2)& xgt-frac 1 2 end{cases}.$$
If $c=0$, for example, $U(0)=1/4$, which is attained by a random variable $X$ taking the value $X=1$ with probability $1/3$ and the value $X=-1/2$ with probability $2/3$. Then $$EX^3=frac 2 3 (-frac 1 8 )+frac 1
3(1)=frac 1 4 .$$
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
|
show 3 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%2f3020531%2fbound-on-third-moment-when-the-variable-is-bounded-and-has-finite-expectation%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
Let $f(x)=x^3$ and let $Ssubset mathbb R^2$ be the convex hull of the points ${(x,f(x)): xin[-1,1]}$. The set $S$ is the set of all possible $(EX,EX^3)$ values for random variables $X$ with $P(|X|le1)=1.$ The intersection of $S$ with the vertical line cut out by $x=c$ is the interval from $(c,L(c))$ to $(c,U(c))$, where $L$ and $U$ are the lower convex and upper concave envelopes of $f$ on $[-1,1]$. Then the desired upper bound is $U(c)$.
More explicitly: Draw a line through $(1,1)$, tangent at $(-1/2, -1/8)$ to the graph of $f$. The graph of $U$ follows that of $f$ for $xin[-1,-1/2]$ and that of the tangent line for $xin[-1/2,1]$. The formula for $U$ is $$U(x)=begin{cases}-x^3&xle -1/2\ -frac 1 8 +frac 3 4 (x+1/2)& xgt-frac 1 2 end{cases}.$$
If $c=0$, for example, $U(0)=1/4$, which is attained by a random variable $X$ taking the value $X=1$ with probability $1/3$ and the value $X=-1/2$ with probability $2/3$. Then $$EX^3=frac 2 3 (-frac 1 8 )+frac 1
3(1)=frac 1 4 .$$
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
|
show 3 more comments
Let $f(x)=x^3$ and let $Ssubset mathbb R^2$ be the convex hull of the points ${(x,f(x)): xin[-1,1]}$. The set $S$ is the set of all possible $(EX,EX^3)$ values for random variables $X$ with $P(|X|le1)=1.$ The intersection of $S$ with the vertical line cut out by $x=c$ is the interval from $(c,L(c))$ to $(c,U(c))$, where $L$ and $U$ are the lower convex and upper concave envelopes of $f$ on $[-1,1]$. Then the desired upper bound is $U(c)$.
More explicitly: Draw a line through $(1,1)$, tangent at $(-1/2, -1/8)$ to the graph of $f$. The graph of $U$ follows that of $f$ for $xin[-1,-1/2]$ and that of the tangent line for $xin[-1/2,1]$. The formula for $U$ is $$U(x)=begin{cases}-x^3&xle -1/2\ -frac 1 8 +frac 3 4 (x+1/2)& xgt-frac 1 2 end{cases}.$$
If $c=0$, for example, $U(0)=1/4$, which is attained by a random variable $X$ taking the value $X=1$ with probability $1/3$ and the value $X=-1/2$ with probability $2/3$. Then $$EX^3=frac 2 3 (-frac 1 8 )+frac 1
3(1)=frac 1 4 .$$
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
|
show 3 more comments
Let $f(x)=x^3$ and let $Ssubset mathbb R^2$ be the convex hull of the points ${(x,f(x)): xin[-1,1]}$. The set $S$ is the set of all possible $(EX,EX^3)$ values for random variables $X$ with $P(|X|le1)=1.$ The intersection of $S$ with the vertical line cut out by $x=c$ is the interval from $(c,L(c))$ to $(c,U(c))$, where $L$ and $U$ are the lower convex and upper concave envelopes of $f$ on $[-1,1]$. Then the desired upper bound is $U(c)$.
More explicitly: Draw a line through $(1,1)$, tangent at $(-1/2, -1/8)$ to the graph of $f$. The graph of $U$ follows that of $f$ for $xin[-1,-1/2]$ and that of the tangent line for $xin[-1/2,1]$. The formula for $U$ is $$U(x)=begin{cases}-x^3&xle -1/2\ -frac 1 8 +frac 3 4 (x+1/2)& xgt-frac 1 2 end{cases}.$$
If $c=0$, for example, $U(0)=1/4$, which is attained by a random variable $X$ taking the value $X=1$ with probability $1/3$ and the value $X=-1/2$ with probability $2/3$. Then $$EX^3=frac 2 3 (-frac 1 8 )+frac 1
3(1)=frac 1 4 .$$
Let $f(x)=x^3$ and let $Ssubset mathbb R^2$ be the convex hull of the points ${(x,f(x)): xin[-1,1]}$. The set $S$ is the set of all possible $(EX,EX^3)$ values for random variables $X$ with $P(|X|le1)=1.$ The intersection of $S$ with the vertical line cut out by $x=c$ is the interval from $(c,L(c))$ to $(c,U(c))$, where $L$ and $U$ are the lower convex and upper concave envelopes of $f$ on $[-1,1]$. Then the desired upper bound is $U(c)$.
More explicitly: Draw a line through $(1,1)$, tangent at $(-1/2, -1/8)$ to the graph of $f$. The graph of $U$ follows that of $f$ for $xin[-1,-1/2]$ and that of the tangent line for $xin[-1/2,1]$. The formula for $U$ is $$U(x)=begin{cases}-x^3&xle -1/2\ -frac 1 8 +frac 3 4 (x+1/2)& xgt-frac 1 2 end{cases}.$$
If $c=0$, for example, $U(0)=1/4$, which is attained by a random variable $X$ taking the value $X=1$ with probability $1/3$ and the value $X=-1/2$ with probability $2/3$. Then $$EX^3=frac 2 3 (-frac 1 8 )+frac 1
3(1)=frac 1 4 .$$
edited Nov 30 '18 at 22:33
answered Nov 30 '18 at 22:00
kimchi loverkimchi lover
9,69631128
9,69631128
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
|
show 3 more comments
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
Why does this property of belonging to the convex hull holds?
– Daniel
Nov 30 '18 at 22:37
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
@Daniel You mean, that $S$ has all possible $(EX,EX^3)$ pairs in it? Intuitively, because the $E$ operator averages, and the convex hull of a set is the set of all possible averages. A high-brow version of this idea is the basis of en.wikipedia.org/wiki/Choquet_theory .
– kimchi lover
Nov 30 '18 at 22:42
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
I never saw this idea, but it seems interesting. I can see easily that $X$ being a constant with probability 1 implies $(c, f(c))$ is in the possible pairs, and thus that any point in the convex hull is possible by considering the random variable $X$ that is assumes value $c_1$ with probability $alpha$ and $c_2$ with probability $1 - alpha$. However I do not see how to ensure that a general random variable is also in $S$. edit I see it now, just consider a sequence of discrete random variables converging to the general one... Thanks!
– Daniel
Nov 30 '18 at 22:57
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
@Daniel The same reasoning should make you believe this for "simple" random variables, taking on finitely many values only. There are 2 parts to my claim: each $(EX,EX^3)$ is a point in $S$, and that each point in $S$ arises this way. Which gives you more trouble?
– kimchi lover
Nov 30 '18 at 23:06
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
A truly clever idea. I have never thought in such way. Thanks!
– Hanae Sakuraba
Nov 30 '18 at 23:10
|
show 3 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%2f3020531%2fbound-on-third-moment-when-the-variable-is-bounded-and-has-finite-expectation%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
Let $P(X=1)=c$ and $P(X=0)=1-c$. Then $E(X)=E(X^3)=c$. I suspect this is optimal. In general you need to know the distribution of $X$.
– herb steinberg
Nov 30 '18 at 21:15