Question $1$:

In a decision tree, what is typically found at a leaf node?

A feature test condition

The median value of a feature

Final prediction or classification

Decision rule for further branching

Explain your answer

Question $2$:

Hard margin SVMs differ from soft margin SVMs in that:

Hard margin SVMs are designed for nonlinearly separable data

Hard margin SVMs allow data points to be misclassified, while soft margin does not

Soft margin SVMs do not consider the margin between classes

Soft margin SVMs allow data points to be misclassified, while hard margin does not

Explain your answer

Question $3$:

Decision trees are prone to overfitting. What is one common way to avoid overfitting in decision trees?

Applying a more complex decision tree algorithm

Regularizing by restricting the maximum depth or other hyperparameters

Allows the tree to adapt freely on the training data

Increasing the maximum depth of tree

Explain your answer

Question $4$:

What is the primary benefit of using the kernel trick compared to adding polynomial features to a model?

The kernel trick avoids the computational complexity of adding a large number of polynomial features.

The kernel trick adds more complexity to the model compared to adding polynomial features.

Polynomial features always result in better model accuracy.

Adding polynomial features significantly improves model interpretability.

Explain your answer

Question $5$:

What is the primary objective of an SVM classifier?

Minimizes the margin between two classes.

Fits as narrow of a street as possible between two classes.

Maximizes the margin between two classes.

Minimizes the margin between two classes.

Explain your answer

Question $6$:

Machine learning models that are intuitive, easy to interpret, and provide simple rules may be called:

Neural networks

Black box models

Brown box models

White box models

Explain your answer