Anomaly Detection on HVCM dataset RFQ subsystem 1.0 (test)
91.7AccuracyPW-First
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PW-FirstSource=ours, Input Channels=14 channels, Subsystems Configuration=4 subsystems pooled2026.05 | 91.7 | 93.3 | 83.5 | 76.4 | 79.4 | 85.4 | |
| PW-First+SESource=ours, Input Channels=14 channels, Subsystems Configuration=4 subsystems pooled2026.05 | 91.7 | 93.8 | 82.5 | 77.3 | 79.5 | 85.8 | |
| StandardSource=ours, Input Channels=14 channels, Subsystems Configuration=4 subsystems pooled2026.05 | 91.5 | 88.2 | 87.8 | 69.2 | 77.1 | 81.9 | |
| DSSource=ours, Input Channels=14 channels, Subsystems Configuration=4 subsystems pooled2026.05 | 91.4 | 90.4 | 87.4 | 70.5 | 77.4 | 82.5 | |
| CNN+LSTMSource=Zhou et al. [5], Learning Paradigm=Supervised, Input Channels=14 channels2026.05 | 90.4 | 93 | 84.4 | 65.5 | 72.1 | 74.4 | |
| KNNSource=Zhou et al. [5], Learning Paradigm=Supervised, Input Channels=14 channels2026.05 | 87.1 | 90 | 75.5 | 57.4 | 64.9 | 66 | |
| SVMSource=Zhou et al. [5], Learning Paradigm=Supervised, Input Channels=14 channels2026.05 | 87.1 | 91 | 93 | 42.5 | 58.1 | 62.9 | |
| LSTM-AESource=Radaideh et al. [4], Learning Paradigm=Unsupervised, Input Channels=RFQ, C-Flux only2026.05 | 87 | 90 | — | — | — | — | |
| RFSource=Zhou et al. [5], Learning Paradigm=Supervised, Input Channels=14 channels2026.05 | 86.8 | 92 | 70.4 | 61.9 | 64.7 | 65.8 | |
| LSTMSource=Zhou et al. [5], Learning Paradigm=Supervised, Input Channels=14 channels2026.05 | 86.3 | 91 | 82.5 | 54.9 | 64.4 | 67.3 | |
| GRU-AESource=Radaideh et al. [4], Learning Paradigm=Unsupervised, Input Channels=RFQ, C-Flux only2026.05 | 85 | 89 | — | — | — | — | |
| ConvLSTM-AESource=Radaideh et al. [4], Learning Paradigm=Unsupervised, Input Channels=RFQ, C-Flux only2026.05 | 85 | 89 | — | — | — | — | |
| CNNSource=Zhou et al. [5], Learning Paradigm=Supervised, Input Channels=14 channels2026.05 | 84.1 | 88 | 91.9 | 38.7 | 53.9 | 59.6 |