23. The densities of U, V, and T: V sample: 100000 V sample: 100000 1.5 1.5 1.0 1.0 0.5 0.5 0.0 0.0 -0.5 0.0 0.5 -0.5 1.0 T sample: 100000 1.0 0.75 0.5 0.25 0.0 0.0 0.5 1.0 -1.0 0.0 1.0 2.0 estimated values for the cdf of T at the values t, i.e., P(T<=t). Note: in the BUGS code, syntax ‘I.2 <- step (0.2-T)’ means I.2 = 1 when 0.2 – T>=0, otherwise I.2 = 0. Hence the cdf of T at the values of t is estimated by nodes I.2, I.4,I.6,I.8,I1.0..I2.0. Here is the statistics of these nodes: node I.2 mean sd 0.02005 0.1402 MC error 2.5% 4.371E-4 0.0 median 97.5% 0.0 0.0 start 1 sample 100000 I.4 0.07908 0.2699 8.451E-4 0.0 0.0 1.0 1 100000 I.6 0.1806 0.3847 0.001211 0.0 0.0 1.0 1 100000 I.8 0.3202 0.4666 0.001496 0.0 0.0 1.0 1 100000 I1.0 0.5017 0.5 0.00163 0.0 1.0 1.0 1 100000 I1.2 0.6805 0.4663 0.001475 0.0 1.0 1.0 1 100000 I1.4 0.8197 0.3845 0.001178 0.0 1.0 1.0 1 100000 I1.6 0.9191 0.2727 8.145E-4 0.0 1.0 1.0 1 100000 I1.8 0.98 0.1398 4.354E-4 1.0 1.0 1.0 1 100000 I2.0 1.0 0.0 3.162E-131.0 1.0 1.0 1 100000 So the estimated values for the cdf of T at these values would be t 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 P(T<=t) 0 0.020 0.09 0.18 0.32 0.50 0.68 0.82 0.92 0.98 1.0 24. i). Consider U~exp(1), V~exp(1), T = U*V*I(U<1,V<1) The BUGS code model { U~dexp(1) V~dexp(1) U.1 <- step(1.0 - U) V.1 <- step(1.0 - V) T<-U.1*V.1 } The densities of U, V and T would be: U sample: 100000 V sample: 100000 1.0 0.75 0.5 0.25 0.0 T sample: 100000 1.0 0.75 0.5 0.25 0.0 0.0 5.0 2.0 1.5 1.0 0.5 0.0 0.0 10.0 5.0 10.0 0.0 20.0 40.0 25. a) The priori distribution is:š(š) = 1, And f(x = k|p) = (šš)šš (1 − š)š−š , where n = 5, k = 0,1,..5 The posterior distribution of p|x ļ¦5 ļ¶ f (p | x ļ½ k) ļµ f(x ļ½ k | p) f(p) ļ½ ļ§ ļ· p k (1 ļ p)5ļk ļµ p(k ļ«1)ļ1 (1 ļ p)(6ļk) ļ1 , where k = 0,1,2,3,4,5 ļØkļø This is Beta distribution (k+1,6-k) b). node p mean 0.4286 sd 0.1749 MC error 2.5% 5.577E-4 0.1187 median 97.5% 0.4209 0.7759 start 1 sample 100000 95% CI for p: (0.1749,0.7759) 26. Summary of Tstar: > summary(Tstar) Min. 1st Qu. Median Mean 3rd Qu. 0.1800 0.6500 0.8600 0.8514 1.0400 # calculate the standard deviation > sd(Tstar) [1] 0.2173325 Max. 2.1200 # same result if apply sd formula > sqrt(sum((Tstar-mean(Tstar))^2)/(10000-1)) [1] 0.2173325 Hence: ET20 =0.8514. VarT20 =0.2173 From the quantile result list, the first deciles is 0.59, and 9th deciles is 1.135 27. summary of statistic of parameter bootstrapping > summary(Tstar) Min. 1st Qu. Median 0.1800 0.6250 0.7750 > sd(Tstar) [1] 0.2479007 Mean 3rd Qu. 0.8035 0.9550 Max. 1.9850 Hence: ET20 =0.8035. VarT20 =0.2479 c) The summary of statistic for mean = 1.0 using parametric bootstrapping is: > summary(Tstar) Min. 1st Qu. Median 0.1650 0.5600 0.6950 Mean 3rd Qu. 0.7174 0.8500 Max. 1.7750 5 0 1 2 3 4 I(x,y) MSE for alpha = 1/2 MSE for alpha = 2/3 0 1 2 mu 3 4