S3U

Science / AI literacy / Grade 8 / sj890

Test a Predictor on Unseen Examples

Calculate a result, inspect mistakes and explain what a small test cannot prove.

All worksheets
5 questions0m 0s
Learn the skill

Training examples help fit a model. Separate validation examples can guide revisions. A final test set should stay out of that process until evaluation. Accuracy is correct predictions divided by all predictions. It summarizes this test, not every future situation.

Worked example

A paper predictor gets 9 of 12 new examples right. Its accuracy on those examples is 9 / 12 x 100 = 75%. If we repeatedly tune it using the final test answers, those answers are no longer an independent final check.

Eight healthy and six damaged leaves are correctly predicted. Four damaged leaves are missed, and two healthy leaves are wrongly flagged.
Accuracy: 14 / 20 = 70%. This summarizes only this test.

Synthetic Leaf Test

On 20 unseen invented leaf cards, the model correctly labels 8 healthy and 6 damaged leaves. It wrongly calls 4 damaged leaves healthy and 2 healthy leaves damaged. The counts describe only these cards, not actual crop safety. No learner photos, names or uploaded work are used.

Question 1 How many predictions are correct?
Question 2 What is accuracy on these 20 cards, as a whole percent?
Question 3 How many damaged leaves were incorrectly called healthy?
Question 4 Which approach preserves an independent final test?
Question 5 What can this result establish?