Reality vs Prediction Confusion Matrix:
- Prediction Yes / Reality Yes: 110 (True Positive)
- Prediction Yes / Reality No: 60 (False Positive)
- Prediction No / Reality Yes: 50 (False Negative)
- Prediction No / Reality No: 410 (True Negative)
Total: 630 tests
An AI model has been developed to test specimens of blood/urine/cough etc. to diagnose ailments (diabetes/liver infection etc.). The model was tested on a data-set of about 630 tests and the resulting confusion matrix is as follows :
Generated by claude-sonnet-4-6 · 2026-06-21 03:19 · grounding stimulus
Model Answer
(A)
True Negative (TN) = 410
These are cases where the model predicted No and the reality was also No.
---
(B)
Given: TP = 110, FP = 60, FN = 50, TN = 410
Precision = TP / (TP + FP) = 110 / (110 + 60) = 110 / 170 ≈ 0.647
Recall = TP / (TP + FN) = 110 / (110 + 50) = 110 / 160 ≈ 0.6875
F1 Score = 2 × (Precision × Recall) / (Precision + Recall)
= 2 × (0.647 × 0.6875) / (0.647 + 0.6875)
= 2 × 0.4448 / 1.3345
≈ 0.667
Source: Confusion Matrix & Model Evaluation, AI Chapter
---
Explanation
- TN is straightforward — read directly from the matrix (Prediction No / Reality No).
- For Precision, use only the "Prediction Yes" row (TP + FP in denominator).
- For Recall, use only the "Reality Yes" column (TP + FN in denominator).
- F1 Score is the harmonic mean of Precision and Recall — always show the formula first, then substitute values. Examiners award marks for correct formula + correct substitution + correct answer.