Anomaly Detection on SMAP (test) benchmark leaderboard · SOTA2 Research

Anomaly Detection on SMAP (test)

94.7PrecisionDFM

Evaluation Results

Method
DFM2024.10	94.7	96.4	98.1	98.7
LightTS2024.11	92.58	69.21	55.27	—
Pyraformer2024.11	92.54	71.09	57.71	—
DLinear2024.11	92.32	69.26	55.41	—
ETSformer2024.11	92.25	69.5	55.75	—
Anomaly2024.11	91.85	71.18	58.11	—
Reformer2024.11	90.91	70.4	57.44	—
PatchTST2024.11	90.64	68.82	55.46	—
GPT4TS2024.11	90.6	72.88	60.95	—
FEDformer2024.11	90.47	70.76	58.1	—
Autoformer2024.11	90.4	71.12	58.62	—
Peri-midFormer2024.11	90.4	68.62	56.1	—
TimesNet2024.11	90.14	69.39	56.4	—
Informer2024.11	90.11	69.92	57.13	—
DFM2024.10	89.7	94.1	98.9	98.6
Transformer2024.11	89.37	69.7	57.12	—
Non-stationary Transformer (Station)2024.11	89.37	71.09	59.02	—
I-CFM2024.10	89.3	93.6	98.2	98.3
LogTrans2024.11	89.15	69.97	57.59	—
ContextFlow++2024.06	88.64	93.62	99.19	98.66
FM2024.10	88.2	93.3	98.9	98.5
OmniAnomaly2026.06	88.2	91	94	—
I-CFM2024.10	88	93.3	99.2	98.6
Base CNF2024.10	87.5	92.8	98.8	98.4
Generalist2024.06	87.4	86.05	84.93	91.55
Glow DNF2024.10	87.4	86.1	84.9	91.6
FM2024.10	87.4	93.1	99.6	98.7
Base CNF2024.10	86.5	89.1	91.9	94.9
CAE-M2024.06	81.93	88.27	95.67	99.01
CAE-M2024.10	81.9	88.3	95.7	99
OmniAnom.2024.06	81.3	87.28	94.19	98.89
OmniAnomaly2024.10	81.3	87.3	94.2	98.9
Single-entity HTM2026.06	81.2	84.1	87.3	—
TranAD2024.06	80.43	89.15	99.99	99.21
TranAD2024.10	80.4	89.2	99.9	99.2
MTAD-GAT2024.06	79.91	88.8	99.91	98.44
GDN2024.06	74.8	85.18	98.91	98.64