Week 5 · Day 2 — Your First ML Model: Decision Trees

Theme: Machine Learning Tryouts 🤖⚽ Time: ~75 minutes Tools: Python, scikit-learn, pandas Goal: Train a decision tree to predict match outcomes (W/D/L)

Warm-Up: How a Decision Tree Thinks

A decision tree is basically a flowchart the computer builds itself.

Imagine a coach's decision process:

Did we have >55% possession?
├── YES → Did we have >5 shots on target?
│         ├── YES → Predict: WIN
│         └── NO  → Predict: DRAW
└── NO  → Did they have >2 corners in last 15 min?
          ├── YES → Predict: LOSS
          └── NO  → Predict: DRAW

The computer builds this tree automatically by scanning thousands of matches and finding which splits separate wins from losses best.