decision_tree.md

decision tree

Language: Python · Sphere: programming · Category: Reasoning

What it does

Simple Decision Tree Classifier

A complete implementation of a decision tree classifier with information gain, prediction, and pruning capabilities.

Guarantee

When it runs, decision tree guarantees preds == truth; sum(preds) == 6; train_acc == 1.0 (proven by run).

Checkable constraints:

• train_acc == 1.0
• preds == truth
• sum(preds) == 6
• acc_after >= acc_before
• acc_after == 1.0
• flat.predict([[9, 9], [0, 0]]) == [7, 7]

Verification evidence

• Green-run: ✓ passes (re-run under the extractor's gate)
• Constraint strength: recovery (truth-pinned)
• Independent oracle: ✓ consensus — xlang (validator v1.9)
• Peer review: unreviewed

---

△ AURA Pattern Library — © Reality Optimizer