I. 'I'ruefFalserncertain - Briey explain. No credit without an explanation [5 marks each]. 1. 2. 3. GMSE is bigger 1with more Ks than the true model than 1with less Ks. If the exit criterion sls in automatic search is elevated, then the model ends up 1with lea Ks in it. If ii = 0.01 in a logit model. then for each unit increase in X. the probability that Y=l is on average higher by 1%, celen's paribas. . A model 1with 93% correct classication rate is a useful model. . With the cumulative logit, the probability to he in the second highest category K-l or higher is PKl = Fir-1 l. . If the p-value=. oftesting Poisson vs. Negative Binomial model with the xi distribution1 one should use the Poisson model. . 'With all observations in the same cluster, ESQ = l. . With 5m multiple imputation samples generating EDD predictions of I] or 1 for the binary dependent variable for each observation, to combine the EDD predictions into- one for each observation. one has to take the mean of all 5m predictions and use a cutoff of 50%