Data Science Supervised Learning Models 5

Question 1

Which activation function is commonly used in hidden layers to avoid vanishing gradients?

Accepted Answer

ReLU

Answer

Sigmoid

Answer

Tanh

Answer

Softmax

Question 2

What does the softmax function produce in a multiclass classifier?

Accepted Answer

A probability distribution over classes

Answer

A single binary output

Answer

A regression value

Answer

A distance metric

Question 3

When would you prefer mean absolute error over mean squared error?

Accepted Answer

When you want robustness to outliers

Answer

When you want to penalize large errors more

Answer

For classification tasks

Answer

To maximize variance

Question 4

What is the purpose of a confusion matrix?

Accepted Answer

Summarize correct and incorrect predictions by class

Answer

Display feature correlations

Answer

Plot the loss curve

Answer

Rank feature importance

Question 5

Which scenario most likely indicates overfitting?

Accepted Answer

High training accuracy but low test accuracy

Answer

Low training and low test accuracy

Answer

High accuracy on both sets

Answer

Equal precision and recall

Question 6

What does hyperparameter tuning via grid search do?

Accepted Answer

Exhaustively evaluates combinations of preset parameter values

Answer

Learns weights via backpropagation

Answer

Selects features automatically

Answer

Reduces dataset dimensionality

Question 7

Why might you use stratified sampling when splitting data?

Accepted Answer

To preserve class proportions across train and test sets

Answer

To shuffle features randomly

Answer

To remove duplicate rows

Answer

To scale numeric features