The maximum number of samples required to split a node

The maximum number of leaf nodes in the tree

The maximum number of features considered at each split

The maximum number of levels from root to the deepest leaf node

Accuracy = 50%, F1 for positive class = 0.50

Accuracy = 95%, F1 for positive class = 0

Accuracy = 95%, F1 for positive class = 0.95

Accuracy = 5%, F1 for positive class = 0

A single train-test split is always sufficient regardless of dataset size

Cross-validation eliminates the need for a separate test set entirely

Cross-validation always produces higher accuracy numbers

Cross-validation uses all data for both training and testing, providing a more reliable and less variable performance estimate on small datasets

The algorithm would skip Student ID automatically because it detects it has too many unique values

Student ID would be treated identically to any other feature

Student ID would have zero Information Gain because it has no predictive power

Student ID would have the highest Information Gain because each split creates pure single-sample leaves

KNN will automatically normalise the features internally

Both features will contribute equally to the distance

The Age feature will dominate because it has fewer unique values

The Salary feature will dominate the distance calculations, effectively ignoring Age

A sick person is correctly classified as sick

A healthy person is correctly classified as healthy

A sick person is incorrectly classified as healthy

A healthy person is incorrectly classified as having the disease

The node becomes a leaf, despite its high impurity

An error is raised because the impurity is too high for a leaf

The node is removed from the tree entirely

The node is split because Gini is above the threshold

Precision — because we want to be sure every positive prediction is correct

Accuracy — because it gives an overall picture of model performance

Recall — because we must catch as many actual cancer cases as possible

Specificity — because we want to correctly identify healthy patients

The classifier performs no better than random guessing

The classifier has a 50% false positive rate

The classifier is perfectly accurate on 50% of the data

The model has moderate performance — acceptable for most tasks

0.80

0.70

0.75

0.67