Sigmoid

What assumption is the foundation of a Naive Bayes classifier?

• All features have equal weights

• Features have a hierarchical relationship

• Features are dependent on each other

• Features are independent given the class label

Which of the following is a likelihood function in the context of Naive Bayes?

• P(data)

• P(hypothesis)

• P(data | hypothesis)

• P(hypothesis | data)

How does Naive Bayes handle continuous data?

• By converting them to categorical variables.

• By assuming a continuous distribution, such as Gaussian.

• By using kernel density estimation.

• Continuous data cannot be used with Naive Bayes.

What is a major advantage of Naive Bayes in handling large feature spaces?

• It inherently supports multiclass classification.

• It can handle irrelevant features well.

• It requires minimal computational resources.

• It does not require feature scaling.

Which version of Naive Bayes would be most appropriate for a dataset where features are binary attributes?

• Multinomial Naive Bayes.

• Gaussian Naive Bayes.

• Bernoulli Naive Bayes.

• Complement Naive Bayes.

Which metric can be used for choosing the best split in a classification problem?

• Variance reduction.

• Gini index.

• Euclidean distance.

• Adjusted R-squared.

When is the Gini index considered to be 0 for a decision tree node?

• When all records belong to one class.

• When records are evenly distributed across different classes.

• When the node is at the maximum allowed depth.

• When no further splits are possible.

In the context of decision trees, what is variance reduction used for?

• To measure feature importance.

• To choose the best split in a regression problem.

• To increase the bias of the model.

• To prune the tree.

Which of these is a valid stopping criteria to prevent overfitting in a decision tree?

• When each leaf node has at least one sample from each class.

• When the addition of new nodes no longer significantly improves the model.

• When the tree reaches a predefined number of nodes.

• When all the features are used up.

What does an increase in entropy signify in a decision tree split?

• A more homogeneous node.

• A more heterogeneous node.

• That the split has perfectly classified the data.

• A reduction in the depth of the tree.