Classification Report

EBMCClassifier — 2000 samples · 4 classes

0.305
0.3047
Balanced Accuracy
Average per-class recall
30.65%
Accuracy
Fraction of all predictions that are correct: (TP+TN) / Total
30.47%
Balanced Acc
Average per-class recall, correcting for class imbalance
0.3047
Precision
Of predicted positives, how many are correct: TP / (TP+FP)
0.3065
Recall
Of actual positives, how many are found: TP / (TP+FN). Also called sensitivity
0.3020
F1 Score
Harmonic mean of precision and recall: 2·P·R / (P+R)
0.0741
MCC
Matthews Correlation Coefficient: balanced measure even with imbalanced classes. Range [-1, 1]
0.0735
Cohen κ
Agreement beyond chance between predictions and true labels. Range [-1, 1]
0.5643
AUC
Area Under the ROC Curve: probability that a random positive ranks above a random negative
1.3717
Log Loss
Negative log-likelihood of predicted probabilities. Lower is better

Confusion Matrix

ROC Curve

Precision-Recall Curve

Per-Class Metrics

Feature Importances (Top 20)

Class Distribution (Test Set)

Learned Structure — EBMCClassifier bayesian_network

Open interactive Bayesian network → Download structure (JSON) ↓

4 nodes · 3 edges

Markov blanket of target: gene_055, gene_061, gene_153

Edges

Class→gene_061
Class→gene_055
Class→gene_153

Per-Class Metrics

Class                 Precision     Recall         F1    Support
----------------------------------------------------------------
low_risk                 0.2982     0.3042     0.3012        503
moderate_risk            0.2938     0.2247     0.2547        485
high_risk                0.3064     0.2611     0.2820        494
critical_risk            0.3194     0.4286     0.3660        518