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 has a dataset with 100 features and 10,000 samples. They want to reduce the number of features while retaining as much variance as possible, to improve model training speed and reduce overfitting. Which technique should they use?

⚠ Common exam trap

A common mix-up: candidates confuse regularization (which reduces overfitting by shrinking coefficients) with dimensionality reduction, or they think feature scaling alone can reduce feature count, when PCA is the correct technique for explicitly reducing the number of features while preserving variance.

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

✓

Principal Component Analysis (PCA)

Principal Component Analysis (PCA) is an unsupervised dimensionality reduction technique that transforms the original features into a new set of orthogonal components, ordered by the amount of variance they capture. By selecting only the top principal components, the data scientist can significantly reduce the feature count (e.g., from 100 to 20) while retaining the majority of the dataset's variance, which directly improves model training speed and reduces overfitting.

Answer analysis

Option-by-option breakdown

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

✗
Feature scaling
Why it's wrong here
Feature scaling methods such as min-max scaling or z-score standardization transform each column's range or distribution to a common scale, but they leave the total number of columns untouched. The dataset still contains all 100 original features after scaling, so this technique addresses units and magnitude, not dimensionality. Because the goal is explicitly to reduce the number of features, feature scaling cannot be the correct solution.
✓
Principal Component Analysis (PCA)
Why this is correct
Principal Component Analysis (PCA) is an unsupervised linear dimensionality reduction technique that computes the eigenvectors of the covariance matrix and sorts them by their eigenvalues, representing the amount of explained variance. Projecting the data onto the top k principal components yields k orthogonal composite features that retain the most variance, directly reducing the feature count from 100 to k. In Azure Machine Learning, the PCA module performs this projection as a preprocessing step, making it the appropriate choice for this scenario.
✗
Regularization
Why it's wrong here
Regularization adds a penalty term to a model's loss function; L1 (Lasso) regularization can shrink some coefficients to exactly zero, which effectively performs feature selection inside a supervised learning algorithm. However, regularization is embedded in the training process and does not by itself output a reduced feature matrix; you would need to manually drop the zero-weight features after fitting. The problem asks for reducing the number of features explicitly, which PCA does directly and independently of any particular model.
✗
Cross-validation
Why it's wrong here
Cross-validation splits the dataset into multiple training and validation folds to estimate how well a model generalizes to unseen data and to tune hyperparameters. It is an evaluation procedure, so it never modifies, removes, or combines the 100 features in the dataset. While cross-validation can be used in a pipeline to assess the benefit of PCA or other reductions, it is not a dimensionality reduction technique and cannot lower the feature count on its own.

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