Machine Learning Quiz Answers With Complete Solutions
Adding interaction features in linear regression will Correct
Answers never increase the training RSS
Assume that the LDA classification error is 0.9 on the test set,
then Correct Answers LDA is worse than always predicting the
most common class
Assume that you have a problem with many features and you
expect that only a few of them to be important.
Lin Reg
Lasso
Forward feature selection
Ridge
KNN Correct Answers Lasso, forward feature
Benefit of pruning decision tree Correct Answers Decrease
variance
Bootstrapping constructs data sets by sampling ________
Correct Answers randomly with replacement
Consider a neural network with a single output node, several
input nodes, and no hidden layers. The single unit in this
network uses a sigmoid activation function. If you train this
network using the cross-entropy objective, which other machine
learning method will make the most similar predictions?
SVM
Feed forward neural net
LDA
, Decision tree trained using CART
linear regression
poisson regression
Recurrent neural net
Random forest
Log reg Correct Answers Log reg
Cubic regression splines are:
continuously differentiable
continuous
piecewise linear
piecewise constant Correct Answers continuously
differentiable, continuous
Data points with high leverage in simple linear regression are
ones that Correct Answers have a very different X (feature)
value from other data points
Decision boundary is linear in the feature space in the following
methods:
SVM with a polynomial kernel
Maximum margin classifier
Single layer neural net
Recurrent neural net
Log reg
LDA Correct Answers Maximum margin classifier, log reg,
LDA
Decision trees can handle qualitative features ______ Correct
Answers always
Adding interaction features in linear regression will Correct
Answers never increase the training RSS
Assume that the LDA classification error is 0.9 on the test set,
then Correct Answers LDA is worse than always predicting the
most common class
Assume that you have a problem with many features and you
expect that only a few of them to be important.
Lin Reg
Lasso
Forward feature selection
Ridge
KNN Correct Answers Lasso, forward feature
Benefit of pruning decision tree Correct Answers Decrease
variance
Bootstrapping constructs data sets by sampling ________
Correct Answers randomly with replacement
Consider a neural network with a single output node, several
input nodes, and no hidden layers. The single unit in this
network uses a sigmoid activation function. If you train this
network using the cross-entropy objective, which other machine
learning method will make the most similar predictions?
SVM
Feed forward neural net
LDA
, Decision tree trained using CART
linear regression
poisson regression
Recurrent neural net
Random forest
Log reg Correct Answers Log reg
Cubic regression splines are:
continuously differentiable
continuous
piecewise linear
piecewise constant Correct Answers continuously
differentiable, continuous
Data points with high leverage in simple linear regression are
ones that Correct Answers have a very different X (feature)
value from other data points
Decision boundary is linear in the feature space in the following
methods:
SVM with a polynomial kernel
Maximum margin classifier
Single layer neural net
Recurrent neural net
Log reg
LDA Correct Answers Maximum margin classifier, log reg,
LDA
Decision trees can handle qualitative features ______ Correct
Answers always
