You can now: Explain why a 30-day expiry query is SQL—not AI. Identify where an

Microsoft Learn: Prepare data for ML.NET Microsoft Learn: ML.NET data transformations Microsoft Learn: Load data

1. Is the feature available at prediction time?2. Does it represent useful business behaviour?3. Is

1. Is the feature available at prediction time?2. Does it represent useful business behaviour?3. Is

DAY 6 — FEATURE ENGINEERING Business problem: Prediction unit: Prediction moment: Eligibility SQL rule: What

You can now: Separate supervised, unsupervised and reinforcement learning. Choose classification for categories and regression

Microsoft Learn: Machine-learning tasks in ML.NET Microsoft Learn: How to choose an ML.NET algorithm Microsoft

Exact rule or query?→ Deterministic C#/SQL Known categorical answer?→ Classification Known numeric answer?→ Regression No

DAY 5 — TYPES OF MACHINE LEARNING 1. SUPERVISED LEARNING Definition: Requires historical label:Yes Classification:Predicts:

You can now: Separate training, validation and final test responsibilities. Explain generalization, overfitting and underfitting.