PlanetMath (more info)
 Math for the people, by the people.
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: High Entry average rating: Very high
[parent] proof of Jensen's inequality (Proof)

We prove an equivalent, more convenient formulation: Let $ X$ be some random variable, and let $ f(x)$ be a convex function (defined at least on a segment containing the range of $ X$). Then the expected value of $ f(X)$ is at least the value of $ f$ at the mean of $ X$:

$\displaystyle \operatorname{\mathbb{E}}[f(X)] \ge f(\operatorname{\mathbb{E}}[X]). $

Indeed, let $ c=\operatorname{\mathbb{E}}[X]$. Since $ f(x)$ is convex, there exists a supporting line for $ f(x)$ at $ c$:

$\displaystyle \varphi(x)=\alpha (x-c) + f(c) $
for some $ \alpha$, and $ \varphi(x)\le f(x)$. Then
$\displaystyle \operatorname{\mathbb{E}}[f(X)] \ge \operatorname{\mathbb{E}}[\varphi(X)] = \operatorname{\mathbb{E}}[\alpha (X-c) + f(c)] = f(c) $
as claimed.



"proof of Jensen's inequality" is owned by Andrea Ambrosio. [ full author list (2) | owner history (1) ]
(view preamble)

View style:


This object's parent.
Log in to rate this entry.
(view current ratings)

Cross-references: line, convex, mean, expected value, range, segment, convex function, random variable, equivalent

This is version 3 of proof of Jensen's inequality, born on 2002-06-06, modified 2006-11-05.
Object id is 3060, canonical name is ProofOfJensensInequality.
Accessed 15882 times total.

Classification:
AMS MSC39B62 (Difference and functional equations :: Functional equations and inequalities :: Functional inequalities, including subadditivity, convexity, etc.)
 26D15 (Real functions :: Inequalities :: Inequalities for sums, series and integrals)

Pending Errata and Addenda
None.
[ View all 2 ]
Discussion
Style: Expand: Order:
forum policy
proof of jensen inequality by tze on 2005-06-30 19:41:11
the most important part of this proof is: tangent line at the graph of the convex function, is less or equal to the convex function. Wouldn't it be nice if you can add a proof of that?

I guess it has to do with the property that the derivative of a convex function is increasing. but i'm not sure.
[ reply | up ]

Interact
post | correct | update request | add example | add (any)