Describe fundamental principles of machine learning on Azure →mediumMultiple ChoiceObjective-mapped

AI-900 Practice Question: Describe fundamental principles of machine learning on Azure

What is 'label imbalance' in a classification dataset and how does it affect model training?

⚠ Common exam trap

Candidates often confuse label imbalance with data quality issues like label noise or annotation errors, leading them to pick options A or C instead of recognizing it as a class distribution problem.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

When one class greatly outnumbers others, causing models to be biased toward the majority class

Label imbalance refers to a situation in classification datasets where one class (the majority class) has significantly more samples than other classes (minority classes). This causes the model to become biased toward predicting the majority class, as it minimizes overall loss by ignoring minority classes, leading to poor generalization and low recall for underrepresented classes.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
When labels in the training data contain spelling errors
Why it's wrong here
Spelling errors in labels are a form of label noise that degrades data quality, but they do not alter the frequency distribution of classes in the training set. Class imbalance specifically refers to a skewed ratio where one class has far more examples than another, not to the internal correctness or cleanliness of individual labels. A dataset with typo-laden labels can still be perfectly balanced, while a clean dataset can be highly imbalanced, so this option describes a data-quality problem rather than label imbalance.
✓
When one class greatly outnumbers others, causing models to be biased toward the majority class
Why this is correct
Class imbalance occurs when one class, such as 'non-fraud,' vastly outnumbers another, such as 'fraud.' A model trained on such data can trivially achieve high accuracy by always predicting the majority class, while failing to recognize the minority class that is often the most important to detect. Mitigations include resampling methods like oversampling the minority class or undersampling the majority class, applying class weights to the loss function, and evaluating with precision, recall, or F1-score instead of raw accuracy.
✗
When training labels are applied inconsistently by different human annotators
Why it's wrong here
Inconsistent labeling across human annotators is called inter-annotator disagreement or annotation noise, indicating that the ground truth itself is unreliable. This is a label quality issue that can confuse training, but the term 'label imbalance' refers to the relative number of examples per class, not to whether separate annotators agree on those labels. Even if all annotators perfectly agree, the dataset could still be imbalanced; conversely, disagreement can occur in a perfectly balanced dataset.
✗
When a model produces predictions that don't match any of the training labels
Why it's wrong here
When a model produces predictions that don't match any of the training labels, it is usually facing out-of-distribution inputs or a poorly calibrated decision boundary—a generalization failure rather than a property of the training data. Label imbalance describes the distribution of labels within the training set, such as 95% examples of class A and 5% of class B. The model's output space may include labels not seen during training, but imbalance concerns the prevalence of known classes, not the existence of novel or unseen outputs.

Go deeper

AI-900 Practice Question: Describe fundamental principles of machine learning on Azure

Why each option matters

Option-by-option breakdown

Related to this question