Classifier Metrics Playground
N · π · TPR · FPR → live
Drag panels by their grip to rearrange the layout.
⠇ Edit Layout
↺ Reset
Grab ⠇ on any panel to reorder
Parameters
1x
Dataset size (N)
Log · 100 → 500k
Prevalence (π)
% positives
TPR (Sensitivity)
True Positive Rate
FPR (1−Spec)
False Positive Rate
Live Equations
1x
Confusion Matrix
1x
TP
FN
FP
TN
Prediction
Derived Metrics
1x
Precision
—
F1
—
Accuracy
—
Bal Acc
—
TNR
—
NPV
—
Presets
1x
Balanced
π=50%
Imbalanced
π=10%
V. Imbalanced
π=1%
Perfect
TPR≈100 FPR≈0
Random
TPR=FPR
Reset
defaults
Counts
1x
Rates
1x
ROC · iso-LR⁺
1x
Precision-Recall
1x