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

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

A data scientist is training a classification model on a dataset with 100 features and only 500 labeled samples. The model achieves 99% accuracy on the training data but only 68% accuracy on a held-out test set, indicating overfitting. Which technique is most appropriate to directly address this problem?

⚠ Common exam trap

A common mix-up: candidates assume more data (Option A) is always the best fix for overfitting, but the question explicitly tests the ability to choose a technique that directly addresses the high-dimensional, low-sample scenario without requiring additional data collection.

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

✓

Reduce the number of features used for training

Reducing the number of features directly combats overfitting by decreasing model complexity and the risk of learning noise from irrelevant or redundant features. With only 500 samples and 100 features, the model has a high variance problem; feature selection or dimensionality reduction (e.g., using Azure Machine Learning's Filter-Based Feature Selection or PCA) simplifies the hypothesis space, improving generalization to the test set.

Answer analysis

Option-by-option breakdown

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

✗
Increase the amount of training data by collecting more samples
Why it's wrong here
More data generally helps reduce overfitting, but this option is about 'collecting more samples' which is not always possible and is not a technique applied to the current dataset. The question asks for a direct technique to address overfitting given the existing setup.
When this WOULD be correct
When a model underfits due to insufficient data (e.g., low accuracy on both training and test sets), increasing training data can improve performance by providing more representative examples.
✓
Reduce the number of features used for training
Why this is correct
Reducing the number of features (e.g., via feature selection or principal component analysis) directly lowers the dimensionality of the input space, which reduces the model's capacity to fit idiosyncratic noise in the training data. With fewer parameters to estimate, the model learns simpler, more generalizable decision boundaries instead of memorizing spurious correlations. This is a standard regularization technique, particularly effective when the number of features exceeds the number of samples, and it often improves test accuracy even if training accuracy drops slightly.
✗
Increase the complexity of the model by adding more layers
Why it's wrong here
Adding more layers increases the model's capacity and parameter count, enabling it to represent increasingly complex functions that overfit the training set by capturing noise rather than underlying signal. In an overfitting scenario, the model is already too flexible; further architectural complexity would amplify variance and worsen the gap between training and validation performance. To remedy overfitting, one should instead simplify the model, regularize it, or apply early stopping, not compound the problem with additional layers.
When this WOULD be correct
If the model were underfitting (e.g., low accuracy on both training and test sets) and the dataset were large enough to support a more complex model, adding layers could help capture more complex patterns.
✗
Train for more epochs
Why it's wrong here
Training for more epochs typically increases the risk of overfitting, as the model will continue to learn noise in the training data. Early stopping is a technique to prevent this, but simply increasing epochs is not appropriate.
When this WOULD be correct
When the model is underfitting (e.g., low training and test accuracy) and training loss is still decreasing, training for more epochs can improve performance. For example, a neural network with high bias may need more epochs to converge.

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.

✓Reduce the number of features used for trainingCorrect answer▾

Why this is correct

Reducing the number of features (e.g., via feature selection or principal component analysis) directly lowers the dimensionality of the input space, which reduces the model's capacity to fit idiosyncratic noise in the training data. With fewer parameters to estimate, the model learns simpler, more generalizable decision boundaries instead of memorizing spurious correlations. This is a standard regularization technique, particularly effective when the number of features exceeds the number of samples, and it often improves test accuracy even if training accuracy drops slightly.

✗Increase the amount of training data by collecting more samplesWrong answer — click to see why▾

Why this is wrong here

The model already overfits with 500 samples; adding more data would help generalization but is not the most direct technique to address overfitting caused by high dimensionality relative to sample size.

★ When this WOULD be the correct answer

When a model underfits due to insufficient data (e.g., low accuracy on both training and test sets), increasing training data can improve performance by providing more representative examples.

Why candidates choose this

Candidates know that more data often reduces overfitting, but they overlook that the question asks for the most direct technique given the specific context of many features and few samples.

✗Increase the complexity of the model by adding more layersWrong answer — click to see why▾

Why this is wrong here

Increasing model complexity by adding more layers would exacerbate overfitting, especially given the small sample size relative to the number of features, making the model even more prone to memorizing noise.

★ When this WOULD be the correct answer

If the model were underfitting (e.g., low accuracy on both training and test sets) and the dataset were large enough to support a more complex model, adding layers could help capture more complex patterns.

Why candidates choose this

Candidates may think that more complex models always improve performance, overlooking the bias-variance tradeoff and the risk of overfitting with limited data.

✗Train for more epochsWrong answer — click to see why▾

Why this is wrong here

Training for more epochs would likely increase overfitting further, as the model already memorizes the training data (99% accuracy vs 68% test accuracy). More epochs exacerbate overfitting rather than reduce it.

★ When this WOULD be the correct answer

When the model is underfitting (e.g., low training and test accuracy) and training loss is still decreasing, training for more epochs can improve performance. For example, a neural network with high bias may need more epochs to converge.

Why candidates choose this

Candidates may think that more training always improves accuracy, not realizing that overfitting is already present and more epochs will worsen generalization.

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