Microsoft Azure AI Fundamentals (AI-900) Practice Exam

Training a regression model typically requires labeled data, as these models learn to predict a continuous output based on input features. Labeled data means that you have a dataset where the outcome variable is known, allowing the model to understand the relationship between input features and the output value.

Unlabeled data lacks this critical information about the outcome, making it difficult for traditional regression models to find the patterns needed to make predictions. Although some advanced techniques may utilize unlabeled data in a more complex way, such as semi-supervised learning or using clustering methods to pre-process data, the fundamental principles of regression as a supervised learning task require labels to gauge error and adjust the model during training.

In summary, a regression model's learning process relies on labeled data to understand how to map input features to a specific output, thus making it impractical to train such models solely with unlabeled data.