This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Concordia University Department of Mathematics & Statistics Formula Sheet 1. For an MA(1) process { X t = Z t + Z t 1 ,t Z } , { Z t } WN(0 , 2 ), X (0) = 2 (1 + 2 ), X (1) = 2 , X ( h ) = 0 if h > 1 For an AR(1) process given by { X t = X t 1 + Z t } , where { Z t } WN(0 , 2 ),   < 1, X ( h ) = 2  h  / (1 2 ), h Z 2. Let { Y t } be a stationary time series with mean 0 and covariance function Y . If j =  j  < , then the time series X t = j = j Y t j is stationary with mean 0 and ACVF X ( h ) = X j = X k = j k Y ( h + k j ) . Further, if { Y t } has spectral density f Y ( ), then the spectral density of { X t } is given by f X ( ) =  ( e i )  2 f Y ( ) where ( e i ) = j = j e ij 3. The best linear predictor of Y in terms of 1 , W = ( W n ,...,W 1 ) is P ( Y  W ) = Y + a ( W W ), where Y = E [ Y ] , W = (...
View
Full
Document
 Spring '11
 theodoro
 Statistics

Click to edit the document details