now the formulas for Mallow's Cp, AIC, BIC (also called SBC), R2, adjusted R2, M SE, and SSE. Know that these are all standard goodness-of-fit measures. Mallow's Cp = ... AIC = ... BIC = ... R2 adj = ... R2 = ... SSE = ... M SE = ... Know how the method of best-subsets works- be able to write down the method in your own words. Know generally how step-wise regression methods work, and that they are use- ful when there are too many variables to apply best-subsets, but can lead to mediocre (or worse) models. Given some data and logistic/binomial regression output from R, be able to use that output to predict probabilities of the response outcome (success/failure) based on combinations of the predictors. For categorical predictors: given some data, be able to write down the model matrix. Remember, if you have a categorical variable that can take on one of f factor levels, you should only use f-1 columns (or dummy variables) in your model matrix- else your model matrix will be overdetermined and inverting XT X will be impossible. Example: Animal V1 V2 Response Dog 9 12 32 Dog 8 1 31 Dog 2 7