🧭 01 — Evaluation Overview

Model evaluation answers:

How well will this model perform on new data?

---

Good Evaluation Requires

• unseen test data
• appropriate metric
• no leakage
• comparison to baseline
• error analysis
• business interpretation

---

Baseline Model

Regression baseline:

from sklearn.dummy import DummyRegressor

Classification baseline:

from sklearn.dummy import DummyClassifier

If your model cannot beat a dummy baseline, it is not useful.

---

Next

➡️ 02-cross-validation