AI-900 Practice Question: Describe fundamental principles of machine learning on Azure
What is the purpose of splitting data into training, validation, and test sets in machine learning?
⚠ Common exam trap
Many exam-takers confuse the purpose of splitting with increasing data quantity or speeding up training, not realizing that the core reason is to obtain an unbiased estimate of model performance on unseen data.
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
✓
To evaluate model performance honestly on data it hasn't seen during training
Splitting data into training, validation, and test sets is essential for honestly evaluating a model's performance on unseen data. The training set teaches the model patterns, the validation set tunes hyperparameters and prevents overfitting, and the test set provides a final, unbiased estimate of how the model will perform on new, real-world data. This separation ensures that the model's accuracy metrics reflect its generalization ability rather than memorization of the training data.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
To increase the total amount of data available for training
Why it's wrong here
Splitting a dataset into training, validation, and test partitions reallocates existing examples; it cannot increase total data volume. In fact, the training portion shrinks relative to using the whole dataset, so the split reduces the data available for model fitting. Increasing data volume would require acquiring new samples or generating synthetic examples, not partitioning the current data. The true purpose of the split is to enable honest evaluation on held-out examples.
- ✓
To evaluate model performance honestly on data it hasn't seen during training
Why this is correct
Holding out validation and test data ensures the model is assessed on examples that were never used to update weights or tune hyperparameters, providing an honest estimate of generalization to new data. If the same data used for training were used for evaluation, the scores would be optimistically inflated due to overfitting and memorization. Therefore, the split exists to measure real-world predictive performance, not to enlarge or transform the dataset.
- ✗
To make training faster by using smaller datasets
Why it's wrong here
While using a smaller training set might incidentally shorten each epoch and reduce compute time, that is not why practitioners create train/validation/test splits. Deliberately discarding data just to make training faster would usually degrade model accuracy, because less data often means more variance and worse generalization. The split methodology is a controlled evaluation technique, not a training-speed optimization, so any runtime savings are a side effect rather than the objective.
- ✗
To comply with data privacy regulations
Why it's wrong here
Data privacy regulations like GDPR or HIPAA impose requirements around consent, anonymization, encryption, and access controls throughout the data lifecycle; they do not concern how a dataset is partitioned for model evaluation. A train/validation/test split does not remove personally identifiable information, and sensitive values can remain present in all partitions, so it provides no privacy protection. Compliance is achieved through separate governance practices, not by splitting data into training and test sets.
Go deeper
Related to this question
Learn chapter
Machine Learning Core Concepts
Key term
Machine learning
Machine learning is a branch of artificial intelligence where computers learn patterns from data to make decisions or predictions without being explicitly programmed for every task.
Key term
Training data
Training data is a set of examples used to teach a machine learning model how to make predictions or decisions.
About these practice questions
This AI-900 question is part of Courseiva's 985-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.