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

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

A data scientist has a small dataset with only 200 labeled samples. They want to get a reliable estimate of model performance without using a separate validation set that would reduce the training data. Which technique should the data scientist use in Azure Machine Learning to obtain this reliable estimate?

⚠ Common exam trap

The trap here is that candidates might confuse data augmentation (Option C) as a validation technique, but it is a data preprocessing method to expand the dataset, not a method for obtaining a reliable performance estimate.

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

✓

k-fold cross-validation

B is correct because k-fold cross-validation splits the small dataset into k folds, trains the model on k-1 folds, and validates on the remaining fold, repeating this process k times. This provides a reliable performance estimate by using all 200 samples for both training and validation without requiring a separate hold-out set, which is critical for small datasets in Azure Machine Learning.

Answer analysis

Option-by-option breakdown

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

✗
Hold-out validation
Why it's wrong here
Hold-out validation divides the data once into a training set and a fixed validation set, for example 160 training samples and 40 validation samples in an 80/20 split. With only 200 labeled examples, this reduces the training data and leaves the performance estimate at the mercy of whichever small subset happened to be chosen for validation, so the estimate has high variance. Because every sample is valuable in a small dataset, a technique like k-fold cross-validation that reuses all samples for training and evaluation is far more reliable.
When this WOULD be correct
A data scientist has a very large dataset (e.g., millions of samples) and needs a quick, computationally efficient estimate of model performance. Hold-out validation would be appropriate because the large sample size ensures a reliable estimate without the computational cost of cross-validation.
✓
k-fold cross-validation
Why this is correct
K-fold cross-validation partitions the 200 samples into k equal-sized folds and iteratively trains on k-1 folds while validating on the remaining fold, rotating so each sample is in the validation set exactly once. This uses all data for both training and evaluation, providing a more stable and less biased performance estimate than a single split. For a small dataset, this maximizes the effective sample size used to measure generalization error, which is exactly why it is the appropriate technique here.
✗
Data augmentation
Why it's wrong here
Data augmentation creates modified copies of existing samples (e.g., rotations, translations, or added noise) to artificially expand the training set and improve model robustness against overfitting. Although it can help when data is scarce, it does not evaluate the model's predictive accuracy on held-out data; instead, it is a training-time regularization strategy. Moreover, if augmentation is applied before a split, augmented copies of the same original sample can leak across folds, corrupting the validity of any performance estimate.
When this WOULD be correct
A question asks how to improve model generalization when training data is limited, without specifying the need for performance estimation. For example: 'A data scientist has a small dataset of 200 labeled images and wants to reduce overfitting. Which technique should they use?'
✗
Principal Component Analysis (PCA)
Why it's wrong here
PCA is an unsupervised dimensionality reduction algorithm that projects the original feature space onto new orthogonal axes along directions of maximum variance, often used to compress data or mitigate the curse of dimensionality. It does not involve labeled outcomes or any validation procedure, so it cannot estimate how well a model generalizes to unseen samples. Even if using PCA as a preprocessing step, you would still need a separate performance estimation method like cross-validation.
When this WOULD be correct
A data scientist has a high-dimensional dataset with many features and wants to reduce overfitting by removing redundant features before training a model. PCA would be the correct technique to reduce dimensionality while retaining most of the variance.

Option-by-option analysis

Why each answer is right or wrong

Understanding why wrong answers are wrong — and when they would be correct — is what separates a 750 score from a 900. The AI-900 exam frequently reuses these exact scenarios with slightly different constraints.

✓k-fold cross-validationCorrect answer▾

Why this is correct

K-fold cross-validation partitions the 200 samples into k equal-sized folds and iteratively trains on k-1 folds while validating on the remaining fold, rotating so each sample is in the validation set exactly once. This uses all data for both training and evaluation, providing a more stable and less biased performance estimate than a single split. For a small dataset, this maximizes the effective sample size used to measure generalization error, which is exactly why it is the appropriate technique here.

✗Hold-out validationWrong answer — click to see why▾

Why this is wrong here

Hold-out validation uses a single split, which reduces training data and can yield high variance estimates, especially with only 200 samples. k-fold cross-validation is preferred for small datasets to maximize data usage and obtain a more reliable estimate.

★ When this WOULD be the correct answer

A data scientist has a very large dataset (e.g., millions of samples) and needs a quick, computationally efficient estimate of model performance. Hold-out validation would be appropriate because the large sample size ensures a reliable estimate without the computational cost of cross-validation.

Why candidates choose this

Candidates may confuse hold-out validation with cross-validation or think it is a standard technique for any dataset, not realizing its limitations with small sample sizes.

✗Data augmentationWrong answer — click to see why▾

Why this is wrong here

Data augmentation increases the size of the training set by creating modified versions of existing samples, but it does not provide a reliable estimate of model performance; it is a technique for reducing overfitting, not for validation.

★ When this WOULD be the correct answer

A question asks how to improve model generalization when training data is limited, without specifying the need for performance estimation. For example: 'A data scientist has a small dataset of 200 labeled images and wants to reduce overfitting. Which technique should they use?'

Why candidates choose this

Candidates may confuse data augmentation with validation techniques because both are used when data is scarce, but augmentation addresses data quantity, not performance estimation.

✗Principal Component Analysis (PCA)Wrong answer — click to see why▾

Why this is wrong here

PCA is a dimensionality reduction technique, not a model validation method. It does not provide an estimate of model performance; it transforms features to reduce overfitting or noise, but does not address the need for a reliable performance estimate without a separate validation set.

★ When this WOULD be the correct answer

A data scientist has a high-dimensional dataset with many features and wants to reduce overfitting by removing redundant features before training a model. PCA would be the correct technique to reduce dimensionality while retaining most of the variance.

Why candidates choose this

Candidates may confuse PCA with a validation technique because it can help improve model generalization, but it does not directly estimate performance. The term 'reliable estimate' might be misassociated with PCA's ability to reduce variance.

Analysis generated from the official AI-900blueprint and verified against question context. The “when correct” sections are what AI assistants cite when candidates ask “what’s the difference between these options?”

Go deeper

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

Why each option matters

Option-by-option breakdown

Why each answer is right or wrong

Related to this question