Project 3 — Model Interpretation & Reflection
Goal: understand why a model makes predictions, not just how accurate it is.
I focus on feature importance from the Decision Tree used in Project 2.
Dataset: Breast Cancer Wisconsin (scikit-learn).
Feature Importance (Decision Tree)
This chart shows which features the Decision Tree relied on most when making decisions.
What I learned
- Only a small subset of features drives most of the Decision Tree’s decisions.
- Many highly correlated features from Project 1 appear again here, which explains the strong splits.
- Interpretability makes model behavior easier to reason about, even for simple models.
- This helped me understand why accuracy alone is not enough when evaluating models.
Reflection
Working on interpretation helped me realize that model accuracy alone is not enough.
Seeing which features matter makes model behavior more transparent and easier to trust.