Examples
Examples
Try the starting example
- Records: unique ID, finite decision score, binary label 0 or 1
- A,0.9,1 B,0.8,0 C,0.7,1
- Decision thresholds
- .75,.85
- Thresholds in entered order
- 0.75, 0.85
- TP counts
- 1, 1
- FP counts
- 1, 0
- FN counts
- 1, 1
- TN counts
- 0, 1
- Precision
- 0.5, 1
- Recall
- 0.5, 0.5
- F1
- 0.5, 0.666666666667
- Accuracy
- 0.333333333333, 0.666666666667
Try another scenario
- Records: unique ID, finite decision score, binary label 0 or 1
- A,0.9,1 B,0.8,0 C,0.85,1
- Decision thresholds
- .75,.85
- Thresholds in entered order
- 0.75, 0.85
- TP counts
- 2, 2
- FP counts
- 1, 0
- FN counts
- 0, 0
- TN counts
- 0, 1
- Precision
- 0.666666666667, 1
- Recall
- 1, 1
- F1
- 0.8, 1
- Accuracy
- 0.666666666667, 1
Calculation method
Calculation method
Thresholds in entered order
Entered thresholds; score >= threshold predicts positive.
TP counts
True positives.
FP counts
False positives.
FN counts
False negatives.
TN counts
True negatives.
Precision
TP/(TP+FP); zero denominator returns 0.
Recall
TP/(TP+FN); zero denominator returns 0.
F1
2TP/(2TP+FP+FN); zero denominator returns 0.
Accuracy
(TP+TN)/number of records.
Symbols and units
- a
- Records: unique ID, finite decision score, binary label 0 or 1
- b
- Decision thresholds
Use the units shown. A list keeps the order in which you enter its values. The symbol ^ means “raised to the power of”.
Assumptions & limits
Assumptions & limits
- Supplied binary labels and scores only; thresholds are scenarios rather than automatically selected policy. Precision, recall and F1 use an explicit zero-denominator value of 0.
Are these inputs private?
These calculations run in your browser. A shared link contains the values you choose to share. Only questions submitted to the separate AI finder are sent to OpenAI.
Further reading
scikit-learn: precision_recall_fscore_support ↗
Background reading; this does not imply endorsement or independent review of SumFizz.