Decision Tree

A decision tree is a supervised learning model that reaches a prediction through a sequence of feature-based splits. Each internal node asks a question, each branch represents a possible answer, and each leaf produces a class or numerical value. Trees are relatively easy to visualize and can model nonlinear relationships without extensive feature scaling. A single deep tree may overfit training data, so practitioners limit its depth, prune branches, or combine many trees in ensemble methods such as random forests and gradient boosting. The apparent clarity of a tree does not remove the need to test for unstable splits, biased features, data leakage, and weak performance on new data.