Examples
Examples
Try the starting example
- Records: unique ID, finite decision score, binary label
- A,0.9,1 B,0.8,0 C,0.7,1
- Binary ROC AUC
- 0.5
- Descending distinct score thresholds
- 0.9, 0.8, 0.7
- False-positive rates at thresholds
- 0, 1, 1
- True-positive rates at thresholds
- 0.5, 0.5, 1
- Positive and negative record counts
- 2, 1
Try another scenario
- Records: unique ID, finite decision score, binary label
- A,0.9,1 B,0.8,0 C,0.85,1
- Binary ROC AUC
- 1
- Descending distinct score thresholds
- 0.9, 0.85, 0.8
- False-positive rates at thresholds
- 0, 0, 1
- True-positive rates at thresholds
- 0.5, 1, 1
- Positive and negative record counts
- 2, 1
Calculation method
Calculation method
Binary ROC AUC
Trapezoid area after moving all tied scores together; ties give half pair credit.
Descending distinct score thresholds
Distinct supplied scores descending.
False-positive rates at thresholds
FP/number of actual negatives after threshold inclusion.
True-positive rates at thresholds
TP/number of actual positives after threshold inclusion.
Positive and negative record counts
Actual positive count, actual negative count.
Symbols and units
- a
- Records: unique ID, finite decision score, binary label
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
- Unweighted binary ROC AUC for supplied higher-is-positive scores; the curve also includes its implicit origin (0,0). No confidence interval, multiclass result or model quality guarantee.
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
Background reading; this does not imply endorsement or independent review of SumFizz.