Decision Tree (classifier | regressor)

Fit a tree, plot it, read the rules; control depth, min_samples_leaf and cost-complexity pruning. Concepts: Splits; Gini / entropy / squared error; depth; pruning; rules; no extrapolation Project A: Depth-3 tree: readable approval rules; compare with the A–D risk tiers from the EDA. Project B: Tree regressor cannot forecast a record-high load above anything in training — see the extrapolation problem.