(a) A realtor is studying housing values in the suburbs of Boston, and has given you a dataset with the following attributes for each house: crime rate in the neighborhood, proximity to the Charles River, number of rooms, house color, age of unit, distance to five Boston employment centers, pupil-teacher ratio by town, and house value (the target variable with values high and low). The realtor would like you to build a classification model that not only performs well, but is also easy to interpret. Between nearest neighbor classifiers, C4.5, and ensemble methods, which approach would you choose and why?

(b) Under which scenarios (describe the characteristics of the problem/data) would you prefer Ripper to C4.5 and vice-versa, and why?