ECME Algorithm Step 1(E-step)Set V()=Z1Zo2In and,fori=0,1 calculate s=E(uTuily)o?=02() 01yTp(Z:ZTP(y+tr(o2I-01ZTV()Z) where p()=V()-1-V(-X(XTV-X)-XTV(- Step 2 (M-step) Partition 0 as 01=(oo,o2)and 02=B as in ECM. CM-step 1 Maximize complete data log likelihood over 0 a2X+”=s9/i=0,1 CM-step 2 Maximize the observed data log likelihood over given=(): B+1)=(xTV+)-X)-1(XTVt+1)-y (Note:This is the WLS estimator of B.) 19ECME Algorithm Step 1 (E-step) Set V (t) = Z1Z 0 1σ 2 (t) 1 + σ 2 (t) 0 In and, for i = 0, 1 calculate sˆ (t) i = E(u T i ui |y) | σ 2 i = σ 2(t) i = σ 4 (t) i y TP (t)ZiZ T i P (t)y + tr(σ 2 (t) i Iqi − σ 4 (t) i Z T i V (t)−1 Zi) where P (t) = V (t)−1 − V (t)−1X(XTV (t)−1X) −XTV (t)−1 Step 2 (M-step) Partition θ as θ1 = (σ 2 0 , σ 2 1 ) and θ2 = β as in ECM. CM-step 1 Maximize complete data log likelihood over θ1 σ 2 (t+1) i = sˆ (t) i /qi i = 0, 1 CM-step 2 Maximize the observed data log likelihood over θ given θ (t) 1 = (σ 2 (t) 0 , σ 2 (t) 0 ): β (t+1) = (XTV (t+1)−1X) −1 (XTV (t+1)−1 )y (Note: This is the WLS estimator of β.) 19