KEMBAR78
April 1st, 2016 - Lecture: Covariance and Chebyshev's Inequality | PDF | Covariance | Variance
Scribd
Upload
70 views3 pages

April 1st, 2016 - Lecture: Covariance and Chebyshev's Inequality

The document summarizes key topics from a lecture on covariance and Chebyshev's inequality: 1) It defines covariance and provides intuition for when random variables have positive, negative, or zero covariance based on whether their values swing together, opposite, or independently. 2) It gives the formula for covariance in terms of expected values and provides an example where covariance is positive and another where it is negative. 3) It introduces Chebyshev's inequality and how it provides tail bounds for a random variable using its expected value and variance.

Uploaded by

shamaas hussain
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
70 views3 pages

April 1st, 2016 - Lecture: Covariance and Chebyshev's Inequality

The document summarizes key topics from a lecture on covariance and Chebyshev's inequality: 1) It defines covariance and provides intuition for when random variables have positive, negative, or zero covariance based on whether their values swing together, opposite, or independently. 2) It gives the formula for covariance in terms of expected values and provides an example where covariance is positive and another where it is negative. 3) It introduces Chebyshev's inequality and how it provides tail bounds for a random variable using its expected value and variance.

Uploaded by

shamaas hussain
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

1.

April 1st, 2016 - Lecture:


Covariance and Chebyshev's
Inequality
1. Topics
Variance of a sum of dependent random variables and the definition of covariance.

Properties of covariance. Independence implies correlation but the converse does not hold.

Statement of Chebyshev's inequality and a basic application.

2. Introduction
If X and Y are indepedant, thenV(X+Y)=V(X)+V(Y)

If not, then equality may or may not hold.

Example: A case where equality does not hold.

o Take P(X=1)=12, P(X=0)=12, x=1Y=0, x=1,Y=1.V(Y)=14

o SoV(X)+V(Y)=14+14=12

o V(X+Y) equals 0? X+Y will always be 1.V(X+Y)<V(X)+V(Y)

3. Variance
"Swinging together"V(X+Y)<V(X)+V(Y)

"Swings unrelated"V(X+Y)=V(X)+V(Y)

"Swings are opposite each other"V(X+Y)>V(X)+V(Y)

We want to measure "how much they're together or opposite."

The covariance of X and Y is defined to be:Cov(x,y)=E[E(xE[x])(yE[y])]


Intuition

o "Swing together"Cov(X,Y)>0

o "Swings unrelated"Cov(X,Y)=0

o "Swings opposite"Cov(X,Y)<0

Example (non rigourous pictures)

o Examples, x is height and y is shoe sizeCov(X,Y)>0

o v | .

o a | . .

o l | . . .

o u | . . .

o e | . . .

o | . . .

o y | . .

o | .

o 0 +-----------------

o 0 value of X

o Examples, x is temp and y is sales of hot chocolateCov(X,Y)<0

o v | .

o a | . .

o l | . . .

o u | . . .
o e | . . .

o | . . .

o y | . .

o | .

o 0 +-----------------

o 0 value of X

Formula:Cov(X,Y)=E(XY)E(X)E(Y)

Claim: If X,Y are indepedant, Cov(X,Y)=0

o ProofCov(X,Y)=E(XY)E(X)E(Y)

=E(X)E(Y)E(X)E(Y)

=0

4. Tchbycheff's Inequility
Recall Markov's give tail bound base only on E(X)

Chebyshev's give tail bound using E(X) and V(X)

Chebyshev's Theorem: Let X be a random variable with E(X)=

o Then for any >0.

Example: Roll a fair die 100 times and let Z be the sum and X1, ..., X100 be
the outcomes.

o What's the probability that Z are within 50 of its mean?

You might also like