A model predicts no positives. How should undefined precision be handled in an evaluation report?
Classic MLLevel: IntermediateUndefined precision policy
A model predicts no positives. How should undefined precision be handled in an evaluation report?
Hints
- The denominator is zero.
- Report support alongside ratios.
Answer
State the zero-division policy explicitly and show predicted support and recall. A library's substituted zero, one or NaN is a reporting convention, not evidence of reliable positive predictions.
Source and licence: scikit-learn 1.7.2 — model evaluation · BSD-3-Clause · answer: Quantoros editorial