STAT 432: Basics of Statistical Learning
Quiz II - Review Questions
Exercise 1
Two models were fit using ridge regression. One with λ = 0.1 and the other with λ = 0.002. Which was fit
using λ = 0.1?
Model A:
## (Intercept) x1 x2 x3 x4 x5
## -0.7618 0.3647 0.1893 1.0853 0.7744 1.3136
Model B:
## (Intercept) x1 x2 x3 x4 x5
## -0.5782 0.3322 0.1869 0.9741 0.6806 1.1916
Exercise 2
Two models were fit using the lasso. One with λ = 0.1 and the other with λ = 0.002. Which was fit using
λ = 0.1?
Model A:
## (Intercept) x1 x2 x3 x4 x5
## 0.2709 0.0000 0.0000 0.5547 0.2724 0.8676
Model B:
## (Intercept) x1 x2 x3 x4 x5
## -0.7445 0.3581 0.1838 1.0772 0.7665 1.3075
Exercise 3
Two models were selected, one using backwards AIC and the other using backwards BIC. Which was found
using backwards BIC?
Model A:
## (Intercept) x5
## 0.5515 1.1272
Model B:
## (Intercept) x3 x4 x5
## -0.4745 1.0845 0.7446 1.3418
1
Exercise 4
Suppose we estimate the regression coefficients in a linear regression model by minimizing
2
n
X p
X p
X
yi − β0 − βj xij subject to βj2 ≤ s.
i=1 j=1 j=1
for a paricular value of s. Draw a plot that indicates how train and test RMSE are affected by the choice of
s. Assume that simply minimizing RSS would result in a model that overfits. Also assume that a model with
only the intercept would underfit.
Exercise 5
Continue with the setup from Exercise 4. Consider using this model to make a prediction. What happens to
variance as s is increased from 0?
Exercise 6
Continue with the setup from Exercise 4. Consider using this model to make a prediction. What happens to
squared bias as s is increased from 0?
2
Exercise 7
A tree model was fit to training data for a classification problem. Below is a sample of the data used, where
survived is the response. We are attempting to predict the survival status of passengers of the Titanic:
pclass survived sex age sibsp parch
1st survived female 29.0000 0 0
1st survived male 0.9167 1 2
1st died female 2.0000 1 2
1st died male 30.0000 1 2
1st died female 25.0000 1 2
yes sex = mal no
age >= 9.5 survived
died sibsp >= 2.5
died survived
Predict the outcome of survived for the following three test passengers:
pclass sex age sibsp parch
1st male 71 0 0
1st female 48 1 0
2nd male 3 1 1
3
Exercise 8
A bagged tree model (using a total of 3 trees) was fit to training data for a classification problem. Below is a
sample of the data used, where survived is the response. We are attempting to predict the survival status of
passengers of the Titanic:
pclass survived sex age sibsp parch
1st survived female 29.0000 0 0
1st survived male 0.9167 1 2
1st died female 2.0000 1 2
1st died male 30.0000 1 2
1st died female 25.0000 1 2
died died
0.38 0.39
100% 100%
yes sex = male no yes sex = male no died
0.40
100%
survived died
yes sex = male no
0.75 0.18
34% 63%
pclass = 3rd parch < 0.5 survived
0.75
34%
survived died pclass = 3rd
0.58 0.41
16% 11%
age >= 20 pclass = 2nd,3rd
died died survived
0.22 0.43 0.97
66% 14% 20%
died died survived survived died died survived survived
0.19 0.45 0.77 0.91 0.13 0.21 0.75 0.74
66% 10% 6% 17% 52% 7% 4% 37%
Predict the outcome of survived, and predicted probability of suriving for the following test passengers:
pclass sex age sibsp parch
1st male 71 0 0
1st female 48 1 0
4
Exercise 9
Consider fitting classification methods to the data in the plot below. For each part below, specify which of
the two methods will perform better.
1.0
O O OO O O O O
OOOOOO OOO O O
OO OO
OO OO O O O O O OO OO OO OOO O O
O
O OO O OO O O
OO O O OO
OOOOOOOOO O
O O OOO O OO
O O
O O O OO
O X
X XX XXX X X XX
X OOO O OO
OO O O OO O X X X XX X X X O
O
X X O OO O
OO O O O
O OXXXX X X X X X XX XX O O OOOOO
X X
0.5
OO X X X X O
OOOOOO O XX X XX XX X X O O OO
OO OOO X XX X X X X X X XXX X
XX X
XX X OOO
O O OOO O X XX XX X X X X X X O
OO
O
XX X X XX X XXX O O O
OO O X X X XXX XX X XXX XX XX X OO OO
OO XX X X XX X XX X X X
OX X XX XX XX X
X X X XXX XXX O O
X X
XX X X X OOOOO
X XX X X
O XXX XX X X X XX XX XXX X X O
0.0
O X X
X2
OO X
X X XXX X X X O
O X X
XX XX XXXX X
XX XXXX XX XO
OO
OO X XXX XX XXXXXX X XXX XXXX XXX OO O
O XX XX X O
O
OOO X XXXXX X X X X X X
X XX X XXX X X O OOOO O
OO
OO OO XXX X XX X
XX XX
X X X XX XX XX X OO
X X
O OO X
X XXX X XXX XX XX XX X X O O OO
O
OO O X XXX
−0.5
O OOO XXX X XX X X X X O O O
X X X
X X XXX X X X X X OO OO O O
OOO OO OOOOO
O
XXX
X
X X X X
O O O
OO O O O X XX X X X X X X
XX X XXXX X O OO O O O
O
O XX XX X
OO O O
OO OO XX X
X XXOO
XX
O OOOO
OO
O O OO OOOOO O
OO O O O OO O O OO O O OOO O O
O OO O OO
O O
O OOO OOO O O O O O OOOO OOO
O OOOO OO O
OOO O O O O
−1.0
O OO
OO O O O O O O OO OO
OO O
OOO OOO O O
O
OO OO O OOO
−1.0 −0.5 0.0 0.5 1.0
X1
(a) Additive Logistic Regression or k-Nearest Neighbors?
(b) Additive Logistic Regression or a single Decision Tree?
(c) Additive Logistic Regression or a Random Forest?
(d) A single decision tree or a Random Forest?