2_Probability and Statistics

9112012 p kolm 17 correlation corrxy covariance

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: E éêë(X - mX )(Y - mY )ùúû VER. 9/11/2012. © P. KOLM 17 Correlation – Corr(X,Y) Covariance is dependent upon the units of X and Y (i.e. Cov(aX ,bY ) = a ⋅ b ⋅ Cov(X ,Y ) ) Correlation, denoted by Corr (X ,Y ) or rXY , “normalizes” the covariance by the standard deviations of X and Y, Corr (X ,Y ) º rXY º sXY Cov(X ,Y ) = sX sY Var (X ) ⋅Var (Y ) We therefore have that -1 £ Corr (X ,Y ) £ 1 VER. 9/11/2012. © P. KOLM 18 More on Correlation and Covariance If sXY = 0 (or equivalently rXY = 0 ) then X and Y are linearly unrelated If rXY = 1 then X and Y are said to be perfectly positively correlated If rXY = -1 then X and Y are said to be perfectly negatively correlated Note that Corr (aX ,bY ) = Corr (X ,Y ) if ab > 0 Corr (aX ,bY ) = -Corr (X ,Y ) if ab < 0 VER. 9/11/2012. © P. KOLM 19 Some Properties of the Expected Value Let X and Y, and a and b, denote two random variables and two constants, respectively. Then, E (a ) = a , Var (a ) = 0 E (E (X )) = E (X ) , i.e. E (mX ) = mX E (aX + b) = aE (X ) + b E (X +Y ) = E (X ) + E (Y ) E ((aX )2 ) = a 2E (X 2 ) VER. 9/11/2012. © P. KOLM 20 Some Properties of the Variance and the Covariance 2 Var (X ) = E (X 2 ) - mX Var (aX + b) = a 2Var (X ) Var (X +Y ) = Var (X ) +Var (Y ) + 2Cov(X ,Y ) Var (X -Y ) = Var (X ) +Var (Y ) - 2Cov(X ,Y ) Cov(X ,Y ) = E (XY ) - mX mY If X and Y are independent, then Cov(X ,Y ) = 0 . (Note: The converse is not true.) (For you: Verify these properties.) VER. 9/11/2012. © P. KOLM 21 The Normal Distribution We denote by N (m, s 2 ) the normal distribution, with mean m and variance s 2 The probability density function (pdf) of the normal distribution is given by æ (x - m)2 ö 1 ÷ ÷ f (x ) = exp çç ÷ 2 ç ç 2s ÷ è ø s 2p VER. 9/11/2012. © P. KOLM 22 The Standard Normal Distribution Any random variable can be “standardized” by (1) subtracting its mean, and (2) dividing by its standard deviation. In other words, by “standardizing” X we X - mX obtain the new random variable i.e. Z = sX o Note that E (Z ) = 0 and Var (Z ) = 1 A standardized normal variable follows a standard normal distribution, denoted by N (0,1) . Its pdf is given by j(z ) = VER. 9/11/2012. © P. KOLM 1 2p æ -z 2 ö ÷ ÷ exp ç ç ç2÷ ÷ ç è ø 23 Some Properties of the Normal If X N (m, s 2 ) then aX + b N (a m + b, a 2s 2 ) A linear combination of independent, identically distributed (iid) normal random variables are normally distributed s2 If Y1,Y2,... n N (m, s ) are iid then E (Y ) N (m, ) Y n 2 VER. 9/11/2012. © P. KOLM 24 The Cumulative Distribution Function For a pdf, f (x ), where f (x ) P(X = x ), the cumulative distribution function (cdf), F (x ) , is P(X £ x ) ; P (X > x ) = 1 - F (x ) = P (X < -x ) The pdf for the standard normal is f(z ) . Its cdf is denoted by F(z ) = P (Z < z ) o P ( Z > a ) = 2P (Z > a ) = 2[1 - F(a )] o P (a £ Z £ b) = F(b) - F(a ) VER. 9/11/2012. © P. KOLM 25 The Chi-Square Distribution The chi-square distribution is obtained directly from independ...
View Full Document

This document was uploaded on 02/17/2014 for the course COURANT G63.2751.0 at NYU.

Ask a homework question - tutors are online