AnomalyNet Hybrid
Trained on European credit card data (Sep 2013, 284,807 transactions). Isolation Forest detects anomalies by how easily a transaction can be isolated. The Autoencoder is trained exclusively on normal transactions — high reconstruction error flags anomalies. A weighted hybrid achieves the best AUC.
Testing Lab
Enter values to run a single prediction
~0.97
Hybrid AUC
~0.88
IsoForest AUC
~0.95
AutoEnc AUC
Unsupervised
Approach
Security Report
Fill in the fields and run analysis to see the prediction result
Architecture Notes
Isolation Forest: 200 estimators, contamination=~0.17%. Autoencoder: Input(31)→Dense(16)→Dense(8)→Bottleneck(4)→Dense(8)→Dense(16)→Output(31). Trained with MSE loss on normal transactions only. Hybrid score = w_iso × IF_score + w_ae × recon_error (weights from AUC).