How m4444sa/tic-tac-toe-tree works
This is a single-notebook machine-learning study that classifies completed Tic-Tac-Toe boards as X wins or not. Its documented workflow loads the UCI dataset, splits it, one-hot encodes board positions, and compares a decision tree, random forest, and XGBoost; decision-tree depth is chosen using validation accuracy. The notebook also manually calculates entropy and information gain. Source excerpts were unavailable, so links reflect README-documented workflow only.
Main entry points: tic-tac-toe-tree.ipynb One-hot encoding
Components 10
Other components
- Notebook user
- UCI endgame data
- Data split
tic-tac-toe-tree.ipynb - One-hot encoding
tic-tac-toe-tree.ipynb - Entropy and gain
tic-tac-toe-tree.ipynb - Decision tree
tic-tac-toe-tree.ipynb - Random forest
tic-tac-toe-tree.ipynb - XGBoost
tic-tac-toe-tree.ipynb - Depth selection
tic-tac-toe-tree.ipynb - Model comparison
tic-tac-toe-tree.ipynb
Connections 11
- Notebook user → Data split: opens notebook
- UCI endgame data → Data split: provides boards
- Data split → One-hot encoding: passes data
- One-hot encoding → Entropy and gain: provides features
- One-hot encoding → Decision tree: provides features
- One-hot encoding → Random forest: provides features
- One-hot encoding → XGBoost: provides features
- Decision tree → Depth selection: tests depths
- Depth selection → Model comparison: provides selected tree
- Random forest → Model comparison: provides model
- XGBoost → Model comparison: provides model