Class 10 AI Unit 3: Evaluating Models
21 theory hours · 4 practical hours · 10 theory marks
What this unit is about
This unit is about checking how good a model is. You split data into training and testing sets, build a confusion matrix (true/false positives and negatives), and calculate accuracy, precision, recall and F1 score, then choose the right metric for the problem and think about bias and transparency.
Learn it with the AI tutor, one minute each
What the unit covers
- What evaluation is and why a model needs it.
- Train-test split.
- Accuracy and error; the confusion matrix; precision, recall and F1 score.
- Choosing the right metric, and ethical concerns: bias, transparency and accuracy.
Key terms
- Accuracy
- (TP + TN) ÷ all predictions: the share the model got right.
- Precision
- TP ÷ (TP + FP): of everything predicted positive, how much really was.
- Recall
- TP ÷ (TP + FN): of all real positives, how many the model caught.
- F1 score
- 2 × precision × recall ÷ (precision + recall): one number that balances both.
- False negative
- The model says no, but the real answer was yes.
Exam-style questions
1. A model has TP = 40, TN = 45, FP = 5, FN = 10. Find its accuracy.
Show answer →
(40 + 45) ÷ 100 = 0.85, so 85%.
2. For the same model, find precision and recall.
Show answer →
Precision = 40 ÷ 45 ≈ 0.89. Recall = 40 ÷ 50 = 0.80.
3. For detecting a disease, which matters more: precision or recall?
Show answer →
Recall, because a false negative means a sick patient is sent home untreated, which costs more than an extra test.
Sources
Unit names, hours and marks checked against the official CBSE 2026-27 curriculum on 1 November 2026. Explanations and questions are written by CogniGuru.