Which model is designed for classification and can also be applied to regression problems?

• A. Decision tree
• B. Support vector machine
• C. Random forest
• D. K-nearest neighbors

Answer: (A)

Decision trees are built by splitting data based on feature thresholds to form leaves that hold outputs. For classification, those leaves hold class labels; for regression, they hold numeric predictions (often the average target value within that leaf). The way splits are chosen changes with the task: impurity measures like Gini or entropy guide classification splits, while minimizing prediction error (such as mean squared error) guides regression splits. Because the same tree-building framework can handle both objectives, this type of model is designed to work for classification and can be applied to regression as well. This dual capability is the essence of the CART approach (Classification And Regression Trees). So, the model that fits both roles is the decision tree.