+
+
+
+
Confusion Matrix Simulator
Real-time classification telemetry. Dial values smoothly, inspect error trade-offs, and trace how precision, recall, and F1-score are derived from ground truth samples.
01 // Preset Benchmarks
Select a benchmark dataset to inspect real-world classification patterns02 // Dataset Population Breakdown
Actual Positives: 50 (50.0%)•Actual Negatives: 50 (50.0%)•Total N: 100
True Positives:45
False Negatives:5
False Positives:10
True Negatives:40
03 // Classification Matrix
Heatmap Density ActivePREDICTED (+)
Model Positive
PREDICTED (−)
Model Negative
TOTAL
Row Sum
ACTUAL (+)PositiveGround Truth
Condition +
TPHit
45.0%True Positive
Actual (+) · Pred (+)
Correct detection of ground truth positive cases.
FNMiss
5.0%False Negative
Actual (+) · Pred (−)
Type II Error: Ground truth missed by model.
P Sum5050%
ACTUAL (−)NegativeGround Truth
Condition −
FPAlarm
10.0%False Positive
Actual (−) · Pred (+)
Type I Error: False alarm on negative sample.
TNReject
40.0%True Negative
Actual (−) · Pred (−)
Correct rejection of inactive negative cases.
N Sum5050%
PRED. SUMMarginals
Pred (+)TP + FP
55Pred (−)FN + TN
45Total N100
04 // Performance Telemetry
Hover to trace formulaAccuracy(TP + TN) / N
85.0%PrecisionTP / (TP + FP)
81.8%RecallTP / (TP + FN)
90.0%SpecificityTN / (TN + FP)
80.0%F1-Score2·P·R / (P + R)
85.7%05 // Diagnostic Analysis
Balanced Performance: The classifier maintains consistent trade-offs across classes with an F1-Score of 85.7% and Accuracy of 85.0%.