Mission: Find a useful threshold
Adjust the threshold, satisfy the target conditions, then submit your decision.
0
0
0
0
| ID | Actual | Score | Prediction | Outcome | Model spread |
|---|
Upload a CSV containing an actual label column and one or more numeric score columns.
TPR = TP / (TP + FN)
FPR = FP / (FP + TN)
Specificity = TN / (TN + FP)
Youden's J = TPR - FPR
AUC = area under the ROC curve
AUC summarizes ranking discrimination, while threshold metrics describe one operating point.
- Generate a dataset or upload prediction scores.
- Compare model curves and AUC values.
- Move the threshold and watch outcomes change.
- Meet the mission conditions, then submit.
- Use feedback to improve your next decision.