nerdexam
Microsoft

AI-900 · Question #171

During the process of Machine Learning, when should you review evaluation metrics.

The correct answer is D. After you test a model on the validation data. Evaluation metrics are critically reviewed after a machine learning model has been trained and tested on validation data to assess its performance and generalization capabilities.

Submitted by lukas.cz· Mar 30, 2026Describe fundamental principles of machine learning on Azure

Question

During the process of Machine Learning, when should you review evaluation metrics.

Options

  • ABefore you train a model.
  • BAfter you clean the data.
  • CBefore you choose the type of model.
  • DAfter you test a model on the validation data.

How the community answered

(24 responses)
  • A
    4% (1)
  • C
    8% (2)
  • D
    88% (21)

Why each option

Evaluation metrics are critically reviewed after a machine learning model has been trained and tested on validation data to assess its performance and generalization capabilities.

ABefore you train a model.

Reviewing evaluation metrics before training a model is premature as there is no model performance to evaluate yet.

BAfter you clean the data.

Reviewing evaluation metrics after cleaning data but before model training or testing is premature, as the model's actual performance has not been established.

CBefore you choose the type of model.

Reviewing evaluation metrics before choosing a model type is premature because there is no trained model whose performance can be evaluated at this stage.

DAfter you test a model on the validation data.Correct

Evaluation metrics are used to quantify a model's performance, generalization ability, and suitability for its task, which can only be accurately assessed after the model has completed training and made predictions on unseen validation or test data.

Concept tested: ML Model Evaluation Lifecycle

Source: https://learn.microsoft.com/en-us/azure/machine-learning/concept-what-is-machine-learning?view=azureml-api-2#how-machine-learning-works

Topics

#ML lifecycle#Model evaluation#Evaluation metrics

Community Discussion

No community discussion yet for this question.

Full AI-900 Practice