Metric Report Template
Model or experiment:
Module:
Date:
Decision context
What action will the model support?
Target and positive class
Target:
Positive class:
Why this class matters:
Baseline
What simple baseline is used for comparison?
Chosen metrics
| Metric | Why it matters for this decision |
|---|---|
Threshold
Chosen threshold:
Reason:
Expected trade-off:
Error review
| Error type | Why it matters | Action |
|---|---|---|
| False positive | ||
| False negative |
Limitation note
What should a reader not overclaim from these metrics?