Skip to content
PlayOwl

Computers & Digital Literacy · Grade 11

How Machine Learning Systems Are Evaluated

A vendor says its leak-detection model for Cedar Valley's water meters is '97% accurate'. Learners find out what that does and does not mean. They build confusion matrices, compute accuracy, precision and recall with the real formulas, see how rare positives make accuracy misleading, trace how a decision threshold trades false positives for false negatives, choose metrics by the cost of each error, and evaluate fairly with separate test data and subgroup results.

  • 4 lessons
  • About 110 minutes
  • Check what you learned: 10 questions
  • Ways to answer: fill in the blanks, choose an answer, trace code, highlight a text, match pairs and type a number
Try it now (no account) One of the five free starter courses Open in PlayOwl

What learners will discover

  • Build and read a confusion matrix from labelled predictions, naming true positives, false positives, false negatives and true negatives in context.
  • Compute accuracy = (TP + TN) ÷ total, precision = TP ÷ (TP + FP) and recall = TP ÷ (TP + FN), interpret each in context, and explain why accuracy misleads when positives are rare.
  • Explain how raising or lowering a decision threshold trades false positives against false negatives, and choose a metric or threshold from the cost of each error type.
  • Evaluate a model fairly by using separate test data that resembles real use and by comparing error rates across subgroups.

Lessons

  1. The confusion matrix 7 steps4 practiceabout 25 min
  2. Accuracy, precision and recall 7 steps4 practiceabout 30 min
  3. Thresholds and the cost of errors 6 steps3 practiceabout 27 min
  4. Fair and honest evaluation 6 steps3 practiceabout 28 min
  5. Check what you learned 10 questions at the end

Keep exploring

More 11th grade computer skills

Guide for grown-ups