Functional performance testing · Question 51 of 60
A team measures a model's accuracy on a sample from the training dataset and reports "96% ±3% at a 95% CL". What should be criticized?
2 answers are correct. Hard
Show answer
Answer
- B The test dataset must be completely independent of the training and validation data
- D The test data must reflect the inputs in real operation well
Explanation
Statistical criteria need a lot of test data that appeared neither in training nor in validation and that reflects the inputs in real operation well. The statement with margin of error and confidence level, on the other hand, is exactly as it should be.