Consider the following dataset that contains 80 cereal products with their dietary characteristics. Variable Name Description

Name Name of cereal

mfr Manufacturer of cereal:

A = American Home Food Products;

G = General Mills

K = Kelloggs

N = Nabisco

P = Post

Q = Quaker Oats

R = Ralston Purina

type Type of the cereal (cold/hot)

calories Calories per serving

protein Grams of protein

fat Grams of fat

sodium Milligrams of sodium

fiber Grams of dietary fiber

carbo Grams of complex carbohydrates

sugars Grams of sugars

potass Milligrams of potassium

vitamins Vitamins and minerals - 0, 25, or 100, indicating the typical percentage of FDA recommended

shelf Display shelf (1, 2, or 3, counting from the floor)

weight Weight in ounces of one serving

cups Number of cups in one serving

rating A rating of the cereals from Consumer Reports

1) Which variable(s) should be "dummify" (transformed into multiple binary variables)? Why?

2) Which variable(s) should be excluded from the analysis? Why?

3) If we were to use protein and fat as predictors with a model that is sensitive to the scale (such as K-Nearest Neighbor), should we standardize them, given that both of them are already on the same scale (grams). Why or why not?

4) If we were to use this dataset to predict the rating of each cereal, what additional variables should be collected to improve the predictive performance?