Financial Auditor
Model Feedback & Improvements
Track model performance, provide feedback, and propose improvements
Model Accuracy
89.3%
Based on reviewed anomalies
Precision
89.6%
156 true positives
Recall
91.2%
Anomalies caught
False Positives
18
9.6% of reviewed
Model Performance Metrics
Detailed breakdown of model accuracy and effectiveness
Confusion Matrix
True Positives:156
False Positives:18
Total Reviewed:187
F1 Score Breakdown
Precision: 89.6%Accuracy of predictions
Recall: 91.2%Coverage of anomalies
F1 Score: 90.4%Harmonic mean
Performance Insights
• Model is performing well with 89.3% accuracy and balanced precision/recall.
• Low false positive rate (9.6%) suggests good specificity in anomaly detection.
• Consider the 3 proposed improvements to further enhance model performance.
Recent Feedback Summary
How auditors are classifying anomaly predictions
True Positives
156
False Positives
18
Partially Correct
10
Incorrect Severity
3