Anomaly Detection On Ucr Anomaly Archive

Métriques

AUC ROC

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	AUC ROC	Paper Title	Repository
LSTMAD	0.6432	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
LSTM-AE	-	AER: Auto-Encoder with Regression for Time Series Anomaly Detection
Autoencoder (AE)	0.58 ±0.01	Is it worth it? Comparing six deep and classical methods for unsupervised anomaly detection in time series
TranAD	0.4599	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
ARIMA	-	AER: Auto-Encoder with Regression for Time Series Anomaly Detection
OFA	0.5699	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
Robust Random Cut Forest (RRCF)	0.56 ± 0.0019	Is it worth it? Comparing six deep and classical methods for unsupervised anomaly detection in time series
TadGAN	-	AER: Auto-Encoder with Regression for Time Series Anomaly Detection
FCVAE	0.7145	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
SRCNN	0.5109	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
LSTM-VAE	-	AER: Auto-Encoder with Regression for Time Series Anomaly Detection
KAN-AD	0.8188 ±0.0041	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
Auto-Encoder with Regression (AER)	-	AER: Auto-Encoder with Regression for Time Series Anomaly Detection
Graph Augmented Normalizing Flows (GANF)	0.63 ±0.009	Is it worth it? Comparing six deep and classical methods for unsupervised anomaly detection in time series
SAND	0.6550	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
LSTM-DT	-	AER: Auto-Encoder with Regression for Time Series Anomaly Detection
KAN	0.7489	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
SubLOF	0.8001	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
TimesNet	0.4536	KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks	-
MERLIN	0.51 ± 0.0	Is it worth it? Comparing six deep and classical methods for unsupervised anomaly detection in time series

0 of 24 row(s) selected.