A Regression Tree is a tree that predicts a continuous outcome.

A Regression Tree is a tree that predicts a continuous outcome.

• A. Predicts a continuous value
• B. Predicts a probability distribution
• C. Predicts a category
• D. Predicts a binary outcome

Answer: (A)

Regression trees are built to forecast numeric, continuous outcomes. As the tree splits the data into regions, each leaf represents a region where the target values are similar, and the prediction at that leaf is typically the average (or another summary statistic) of the target values in that region. This yields a numeric value for each new observation, which is the essence of predicting a continuous outcome. In contrast, predicting a category is what classification trees do, where the output is a discrete class label. If you’re after a probability distribution over classes, that’s usually produced by probabilistic classifiers or by looking at class probabilities at the leaves of a classification tree, not by a standard regression tree. Predicting a binary outcome is simply a special case of classification, not regression. So the option describing a continuous value best matches what a regression tree actually outputs.

Regression trees are built to forecast numeric, continuous outcomes. As the tree splits the data into regions, each leaf represents a region where the target values are similar, and the prediction at that leaf is typically the average (or another summary statistic) of the target values in that region. This yields a numeric value for each new observation, which is the essence of predicting a continuous outcome.

In contrast, predicting a category is what classification trees do, where the output is a discrete class label. If you’re after a probability distribution over classes, that’s usually produced by probabilistic classifiers or by looking at class probabilities at the leaves of a classification tree, not by a standard regression tree. Predicting a binary outcome is simply a special case of classification, not regression.

So the option describing a continuous value best matches what a regression tree actually outputs.