Homework 2
1. In this problem we use the abalone dataset available on Canvas. The dataset
is about predicting the age of the abalone from its physical measurements. Use the first
7 variables as predictors and the 8-th as the response.
Report all results as the average of 10 random splits with 80% of data for training
and 20% for testing.
a) OLS regression, analytic, by solving the normal equations, with λ = 0.0001.
Report the average training and test R2
.(2 points)
b) Regression tree of maximum depth 1, 2, .... up to 7, for a total of 7 regression
trees. Plot the average training and test R2 vs the tree depth. (2 points)
c) Random forest regression with 10, 30 and 100 trees. Report the average training
and test R2
in each case. (3 points)
1
