Practice Library
All MCQs
Browse exam-wise, subject-wise, and country-wise MCQs with explanations.
Choose an option to check your answer.
A.
The probability of misclassification when labels follow the node's class distribution
B.
The average distance to a centroid
C.
The confidence of an association rule
D.
The number of missing values
Show Answer
Correct Answer: A. The probability of misclassification when labels follow the node's class distribution
Explanation:
Gini impurity is zero for a pure node.
Choose an option to check your answer.
A.
The increase in tree depth
B.
The number of correctly classified training cases
C.
The support of a feature
D.
The reduction in entropy produced by a split
Show Answer
Correct Answer: D. The reduction in entropy produced by a split
Explanation:
Information gain compares parent impurity with weighted child impurity.
Choose an option to check your answer.
A.
The number of tree levels
B.
The average feature value
C.
A measure of class impurity or uncertainty
D.
The distance between leaves
Show Answer
Correct Answer: C. A measure of class impurity or uncertainty
Explanation:
Entropy is low when a node contains mostly one class.
Choose an option to check your answer.
A.
A feature split
B.
A predicted class or class distribution
C.
A support threshold
D.
A missing-value mechanism
Show Answer
Correct Answer: B. A predicted class or class distribution
Explanation:
Leaves terminate decision paths and provide predictions.
Choose an option to check your answer.
A.
A test on an attribute or feature
B.
A final class prediction only
C.
A transaction itemset
D.
A cluster centroid
Show Answer
Correct Answer: A. A test on an attribute or feature
Explanation:
Internal nodes split observations according to feature conditions.
Choose an option to check your answer.
A.
Time data cannot have labels
B.
Cross-validation requires equal feature means
C.
Temporal folds remove all drift
D.
Random folds can expose future information to earlier training periods
Show Answer
Correct Answer: D. Random folds can expose future information to earlier training periods
Explanation:
Forecast evaluation should mimic predicting the future from the past.
Choose an option to check your answer.
A.
Training k models on the same full dataset
B.
Using k nearest neighbors only
C.
Repeatedly training on k-1 folds and validating on the remaining fold
D.
Testing on the training set k times
Show Answer
Correct Answer: C. Repeatedly training on k-1 folds and validating on the remaining fold
Explanation:
Every fold serves once as validation data.
Choose an option to check your answer.
A.
Sorting observations by feature value
B.
Preserving approximately the same class proportions in each split
C.
Using only one class for training
D.
Removing rare classes
Show Answer
Correct Answer: B. Preserving approximately the same class proportions in each split
Explanation:
Stratification reduces accidental class imbalance between partitions.
Choose an option to check your answer.
A.
It focuses on positive-case retrieval and is sensitive to false positives
B.
It ignores positive cases
C.
It requires balanced classes
D.
It always gives higher values
Show Answer
Correct Answer: A. It focuses on positive-case retrieval and is sensitive to false positives
Explanation:
ROC curves can appear optimistic when true negatives dominate.
Choose an option to check your answer.
A.
The accuracy at one fixed threshold
B.
The proportion of positive cases in the sample
C.
The calibration error only
D.
The probability that a random positive receives a higher score than a random negative
Show Answer
Correct Answer: D. The probability that a random positive receives a higher score than a random negative
Explanation:
AUC measures ranking discrimination over all thresholds.
Choose an option to check your answer.
A.
Precision against recall only
B.
Training error against tree depth
C.
True positive rate against false positive rate across thresholds
D.
Support against confidence
Show Answer
Correct Answer: C. True positive rate against false positive rate across thresholds
Explanation:
The ROC curve summarizes discrimination as the decision threshold changes.
Choose an option to check your answer.
A.
When classes are perfectly balanced
B.
When one class greatly outnumbers the other
C.
When every prediction is probabilistic
D.
When there are only two features
Show Answer
Correct Answer: B. When one class greatly outnumbers the other
Explanation:
A model can achieve high accuracy by predicting only the majority class.
