Financial Auditor
Model Test Cases
Test the anomaly detection model with predefined scenarios and custom datasets
Total Test Cases
6
Predefined scenarios
Tests Run
0
In this session
Pass Rate
0%
Tests passed
Large Transaction Detection
Test detection of unusual large amount transactions exceeding $10,000
Transactions
3
Expected Anomalies
1
Anomaly Types
1
Duplicate Transaction Detection
Test detection of duplicate transactions with same amount, date, and merchant
Transactions
3
Expected Anomalies
3
Anomaly Types
1
Unknown Vendor Detection
Test detection of unknown or suspicious vendor transactions
Transactions
3
Expected Anomalies
2
Anomaly Types
1
Combined Anomalies
Test detection of multiple anomaly types in one dataset
Transactions
4
Expected Anomalies
3
Anomaly Types
3
Critical Amount Threshold
Test that amounts over $30,000 are marked as critical severity
Transactions
3
Expected Anomalies
2
Anomaly Types
1
Clean Dataset
Test that a clean dataset with normal transactions produces minimal anomalies
Transactions
3
Expected Anomalies
0
Anomaly Types
0
Detection Model Overview
How the anomaly detection model works
Rule-Based Detection
- • Unusual amounts > $10,000
- • Duplicate transactions
- • Unknown vendors
ML-Based Detection
- • Isolation Forest algorithm
- • Feature engineering pipeline
- • Statistical outlier detection
Test Coverage
These test cases cover:
- ✓ Large transaction detection with severity scaling
- ✓ Duplicate transaction identification
- ✓ Unknown vendor classification
- ✓ Combined anomaly scenarios
- ✓ Confidence score accuracy
- ✓ Clean dataset baseline