Examples Examples Outline 1 Examples Arthur Berg Examples 2/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. Arthur Berg Examples 3/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. var(Xt + Yt ) = var(Xt ) + var(Yt ) < ∞ Arthur Berg Examples 3/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. var(Xt + Yt ) = var(Xt ) + var(Yt ) < ∞ E[Xt + Yt ] = µX + µY Arthur Berg Examples 3/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. var(Xt + Yt ) = var(Xt ) + var(Yt ) < ∞ E[Xt + Yt ] = µX + µY γx+y (h) = cov(Xt+h + Yt+h , Xt + Yt ) Arthur Berg Examples 3/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. var(Xt + Yt ) = var(Xt ) + var(Yt ) < ∞ E[Xt + Yt ] = µX + µY γx+y (h) = cov(Xt+h + Yt+h , Xt + Yt ) = cov(Xt+h , Xt ) + cov(Yt+h , Yt ) + 0 Arthur Berg Examples 3/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. var(Xt + Yt ) = var(Xt ) + var(Yt ) < ∞ E[Xt + Yt ] = µX + µY γx+y (h) = cov(Xt+h + Yt+h , Xt + Yt ) = cov(Xt+h , Xt ) + cov(Yt+h , Yt ) + 0 = γx (h) + γy (h) Arthur Berg Examples 3/ 4 Examples Now your turn! Sum of Uncorrelated Stationary Processes If {Xt } and {Yt } are uncorrelated stationary processes, i.e. if Xs and Yt are uncorrelated for every s and t, show that {Xt + Yt } is stationary and compute its autocovariance function in terms of the autocovariance functions of {Xt } and {Yt }. var(Xt + Yt ) = var(Xt ) + var(Yt ) < ∞ E[Xt + Yt ] = µX + µY γx+y (h) = cov(Xt+h + Yt+h , Xt + Yt ) = cov(Xt+h , Xt ) + cov(Yt+h , Yt ) + 0 = γx (h) + γy (h) Since the mean and autocovariance functions are free of t, the process {Xt + Yt } is stationary. Arthur Berg Examples 3/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. Arthur Berg Examples 4/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. (a) Var[Zt ] = σ 2 (b2 + c2 ) < ∞, E[Zt ] = a, and 2 2 2 σ (b + c ) if h = 0 γZ (h) = bcσ 2 if |h| = 1 0 else is independent of t. Arthur Berg Examples 4/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. (b) Var[Zt ] = b2 σ 2 < ∞, E[Zt ] = a, and γX (h) = b2 σ 2 is independent of t. Arthur Berg Examples 4/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. (c) Var[Zt ] = σ 2 (cos2 (ct) + sin2 (ct)) = σ 2 , E[Zt ] = 0, and γZ (h) = cov(w1 cos(c(t + h)) + w2 sin(c(t + h)), w1 cos(ct) + w2 sin(ct)) = σ 2 (cos(ct) cos(c(t + h)) + sin(ct) sin(c(t + h))) = σ 2 (cos(c(t + h) − ct)) note: cos(x − y) = cos x cos y + sin x sin y = σ 2 cos(ch) is independent of t. Arthur Berg Examples 4/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. (d) Not stationary! var(Zt ) = cos2 (ct) So, for example, var(Z0 ) 6= var(Z1 ). Arthur Berg Examples 4/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. (e) Var[Zt ] = E[Zt Zt−1 ]2 = σ 4 < ∞ E[Zt ] = 0 and γZ (h) = cov(Zt+h Zt+h−1 , Zt Zt−1 ) = 0 is independent of t. In particular, this is an example of a dependent white noise process. Arthur Berg Examples 4/ 4 Examples Now your turn! Stationary? Let wt ∼ iid N (0, σ 2 ), t ∈ Z and a, b, c ∈ (0, 1). Which of the following processes are weakly stationary? For each stationary process specify the mean and autocovariance function. (a) Zt = a + bwt + cwt−1 (d) Zt = w0 cos(ct) (b) Zt = a + bw0 (e) Zt = wt wt−1 (c) Zt = w1 cos(ct) + w2 sin(ct) (f) Zt = wt cos(ct)+wt−1 sin(ct) Note: you should show Var[Zt ] < ∞, and E[Zt ] and γZ (h) are independent of t. (f) var(Zt ) = σ 2 (cos2 (ct) + sin2 (ct)) = σ 2 < ∞. Computing the autocovariance function at h = 1 gives cov(wt cos(ct) + wt−1 sin(ct), wt+1 cos(c(t + 1)) + wt sin(c(t + 1))) = cov(wt cos(ct), wt sin(c(t + 1))) = cos(ct) sin(c(t + 1)) When t = 0, we get cov(Z0 , Z1 ) = sin(c). But when t = −1, we get cov(Z−1 , Z0 ) = 0 Therefore Zt is not stationary! Arthur Berg Examples 4/ 4