Data-Clinic Lab
IN THIS LESSON
Learning Focus: Applying knowledge by training a simple classifier and evaluating its performance.
Essential Question: How do we measure if our AI is any good?
Training & Testing the Classifier
Your goal is to create a simple 'Donut Sorter' that can look at any donut and classify it as either a 'Filled Donut' or a 'Donut with a Hole.' You won't be guessing. You'll act like a machine by first learning from examples and then creating a precise set of rules—an algorithm—to do the job.
🍩 Donut Classifier Training Lab 🤖
Learn Machine Learning by Training Your Own Donut Classifier!
Step 1 of 4
🎯 Lab Setup & Data Exploration
Explore your dataset of 30 donuts and get ready to build your classifier!
Your Mission:
You have 30 different donuts to work with. Your job is to train a classifier that can tell the difference between "Filled Donuts" (🥟) and "Donuts with Holes" (🍩). Take a moment to explore your data!
Record your results for this as you will need them for the reflection lesson next.
Evaluate Your Model
How accurate were your rules? Why did you get some wrong?
Was there a specific type of donut that tricked your rules? (e.g., a powdered donut where the filling mark is hidden).
How could you make your rules even better? What new rules would you add?
This exercise powerfully demonstrates that an AI model is only as good as the rules it's given and the data it's trained on.