Data Science Supervised Learning Algorithms 5

Question 1

What distinguishes a generative classifier like Naive Bayes from a discriminative one like logistic regression?

Accepted Answer

Generative models the joint distribution P(x,y); discriminative models P(y|x) directly

Answer

They are identical approaches

Answer

Discriminative models require more assumptions

Answer

Generative models cannot classify

Question 2

Why might one-hot encoding be necessary before applying linear models to categorical data?

Accepted Answer

Because models treat numeric category codes as having ordinal magnitude

Answer

To reduce dimensionality

Answer

To remove missing values

Answer

To normalize the target variable

Question 3

What is the role of support vectors in an SVM?

Accepted Answer

They are the points closest to the decision boundary that define the margin

Answer

They are the noisy data points removed before training

Answer

They are the centroids of each class

Answer

They are randomly selected samples

Question 4

When predicting a continuous target, which supervised algorithm is appropriate?

Accepted Answer

Linear regression

Answer

Logistic regression

Answer

Naive Bayes classifier

Answer

k-NN classifier

Question 5

What does early stopping accomplish when training a boosted model?

Accepted Answer

It halts adding trees when validation performance stops improving to prevent overfitting

Answer

It removes features

Answer

It increases the learning rate over time

Answer

It guarantees zero training error

Question 6

Which statement about precision is correct?

Accepted Answer

It measures the fraction of predicted positives that are correct

Answer

It measures the fraction of actual positives detected

Answer

It equals recall always

Answer

It ignores false positives

Question 7

In ensemble learning, why does combining diverse weak learners often improve accuracy?

Accepted Answer

Errors of individual learners tend to cancel out when aggregated

Answer

Weak learners never make mistakes

Answer

It reduces the dataset size

Answer

It eliminates the need for validation