Decision Tree Classifier
- Tax Data -

• Before Data Transformation Snapshot

• Tabular data listing tax attributes by State, Group, Item, and Value.

• Includes mixed formats (%, $, numbers) requiring cleaning before analysis.

• Each row represents one tax metric for a state-category (e.g., Colorado's Individual Taxes).

• After Data Transformation Snapshot

• Features (X): Tax indicators by state and group

• Target (y): Binary tax burden (1=High, 0=Low) based on median split

• Only one record is labeled as Low Tax, the rest as High Tax

• Results

• Best Parameters: max_depth=2, min_samples_split=2, criterion='gini'

• Performance: Perfect accuracy (1.00) on training data

• Prediction: All actual and predicted labels match exactly

✅ Conclusion: The model fits perfectly but likely overfits due to small data and class imbalance (only 1 Low Tax sample).