STAT 599 Spring 2013 Homework 6 -------------------------------- For each of these problems, try at least two smoothing methods, trying several choices of the bandwidth or smoothing parameter for each method. Pick a final smooth fit and justify your choice with reasons. 1. Do problem 3 given on page 230 of the book. For this problem, in addition to the nonparametric regression, fit the following nonlinear regression function using the 'nls' function in R: E(Y) = beta0 + beta1*age + beta2*exp(beta3*age) where Y = lweight. Interpret what the estimated parameters tell you and perform any inferences of interest. 2. Do problem 4 given on page 230 of the book. 3. Do problem 5 given on page 230 of the book. NOTE: Examine all relevant aspects of model building that we have discussed in class, such a goodness-of-fit, model diagnostics, etc. For each problem, write a couple of paragraphs summarizing your model and your validation of it. *** For these problems, you will have to load the 'faraway' package in R to access the data sets. If you have never installed it, type install.packages("faraway") and follow the directions. To load the package (and the sm package if necessary) for use in an R session, type library(faraway); library(sm) Make sure you have installed the 'sm' packages from the internet first. Then you can type help(prostate) and help(divusa) and help(aatemp) for details on these data sets.