88%(16)14 out of 16 people found this document helpful
This preview shows page 9 - 14 out of 17 pages.
(e) Check for inﬂuential points.Check the influential points by the half-normal plot on the Cook’s distances.
The point (32) is the most influential point.(f) Check the structure of the relationship between the predictors and the response.5. Using lmod1:(a) Compute and comment on the correlations between the predictors.round(cor(prostate[, -9]), 2) 0.43 0.05 0.28 0.08 0.46 0.63 0.75 1.00(b) Compute the variance inﬂation factors. Should we be worried about multicollinearity? Why/Why not?svi
6. Run stepwise regression on lmod1. Use direction = “both” as one of the parameters of the function. Let us call thislmod1_step. What is the AIC value of lmod1_step? Do you see any reduction in the number of predictors being used inlmod1_step, as compared to lmod1? If yes, what predictors are absent now? What is the Rof the model? Is it better than that oflmod1? Are there still any predictors that are not significant at 95% level? Compare both the models and comment on which ofthem should be preferred and why?