Remaining Useful Life prediction on C-MAPSS FD003
7.18RMSEDMHA-ATCN
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DMHA-ATCNYear=2024, Pre-processing steps=Hybrid CNN2026.02 | 7.18 | — | — | — | — | |
| LSTM (with auto piece-wise)Year=2022, Pre-processing steps=Corr. analysis; median filter; norm; auto piece-wise RUL2026.02 | 8.03 | — | — | — | — | |
| MGLSNYear=2023, Pre-processing steps=Multi-granularity; norm2026.02 | 10.65 | — | — | — | — | |
| HSMGNN2025.12 | 10.88 | 8.5 | 118.27 | — | — | |
| SAGDFN2025.12 | 11.1 | 6.92 | 123.43 | — | — | |
| TATFA-TransformerYear=2024, Pre-processing steps=Transformer + attention2026.02 | 11.23 | — | — | — | — | |
| GA+PredictorYear=2023, Pre-processing steps=GA tuning; norm2026.02 | 11.35 | — | — | — | — | |
| MSAN2026.03 | 11.4 | — | — | 203 | — | |
| BLS+TCNYear=2022, Pre-processing steps=Feature sel.; norm; piece-wise RUL2026.02 | 11.43 | — | — | — | — | |
| GA–RNN/LSTMYear=2021, Pre-processing steps=GA tuning; norm2026.02 | 11.47 | — | — | — | — | |
| FCSTGNN2025.12 | 11.52 | 8.16 | 132.71 | — | — | |
| MsFormer2026.03 | 11.62 | — | — | 236.91 | — | |
| Multi-Scale CNNYear=2020, Pre-processing steps=Multi-scale conv; norm2026.02 | 11.67 | — | — | — | — | |
| DFormerYear=2024, Pre-processing steps=Transformer variant2026.02 | 11.67 | — | — | — | — | |
| MLEAN2026.03 | 11.73 | — | — | 250 | — | |
| DiffRULYear=2024, Pre-processing steps=Diffusion-based RUL2026.02 | 11.77 | — | — | — | — | |
| CNN-SSEYear=2024, Pre-processing steps=CNN + squeeze-excite2026.02 | 11.79 | — | — | — | — | |
| DVGTformer2026.03 | 11.89 | — | — | 254.55 | — | |
| TS-MLLM2026.03 | 11.97 | — | — | 338.3 | — | |
| NSD-TGTN2026.03 | 12.01 | — | — | 220 | — | |
| CDSGYear=2023, Pre-processing steps=Graph structure; norm2026.02 | 12.03 | — | — | — | — | |
| TF-SCN2026.03 | 12.11 | — | — | 238 | — | |
| LSTM + multi-layer self-attnYear=2021, Pre-processing steps=Windows; norm; self-attn2026.02 | 12.13 | — | — | — | — | |
| MAGNN2025.12 | 12.15 | 8.75 | 147.62 | — | — | |
| MegaCRN2025.12 | 12.17 | 8.09 | 148.31 | — | — | |
| PE-Net2026.03 | 12.33 | — | — | 272.85 | — | |
| IMDSSN2026.03 | 12.35 | — | — | 229.54 | — | |
| DeeBERT2026.03 | 12.36 | — | — | 254.19 | — | |
| One Fits AllBackbone=GPT-22026.03 | 12.36 | — | — | 236.43 | — | |
| Bi-level LSTMYear=2022, Pre-processing steps=Hierarchical LSTM2026.02 | 12.37 | — | — | — | — | |
| BiGRU + Temporal AttnYear=2022, Pre-processing steps=Norm; temporal attn2026.02 | 12.45 | — | — | — | — | |
| One Fits AllBackbone=Qwen3-0.6B2026.03 | 12.45 | — | — | 283.98 | — | |
| Variational EncodingYear=2022, Pre-processing steps=Variational reg.; norm2026.02 | 12.51 | — | — | — | — | |
| EAPNYear=2023, Pre-processing steps=Embedded attention; norm2026.02 | 12.52 | — | — | — | — | |
| EAPN2026.03 | 12.52 | — | — | 266.69 | — | |
| Neural ODEYear=2023, Pre-processing steps=ODE layer; norm2026.02 | 12.56 | — | — | — | — | |
| FedLSTMYear=2022, Pre-processing steps=Federated learning; norm2026.02 | 12.57 | — | — | — | — | |
| DTW-GPR2026.03 | 12.7 | — | — | 394.25 | — | |
| Hybrid DL prognosticsYear=2021, Pre-processing steps=Fusion; norm2026.02 | 12.82 | — | — | — | — | |
| Multi-head CNN+LSTMYear=2020, Pre-processing steps=Feature sel.; RUL target2026.02 | 12.85 | — | — | — | — | |
| KDnet2026.03 | 12.95 | — | — | 327.27 | — | |
| Attention-DCNNYear=2021, Pre-processing steps=Conv features; attention; norm2026.02 | 13.08 | — | — | — | — | |
| Crossformer2025.12 | 13.29 | 9.79 | 178.3 | — | — | |
| AGCNNYear=2020, Pre-processing steps=Feature sel.; norm; RUL target2026.02 | 13.39 | — | — | — | — | |
| MSTSDNYear=2024, Pre-processing steps=Multi-scale two-stream; norm2026.02 | 13.66 | — | — | — | — | |
| BiLSTMYear=2018, Pre-processing steps=Feature sel.; norm; RUL target2026.02 | 13.74 | — | — | — | — | |
| BiLSTM2026.03 | 13.74 | — | — | 317 | — | |
| AMR-Net2026.03 | 13.88 | — | — | 372.05 | — | |
| BiGRU2026.03 | 14.02 | — | — | 663.01 | — | |
| Bi-LSTM based Attention methodYear=2022, Pre-processing steps=RUL target function2026.02 | 14.36 | — | — | — | — | |
| GAT-DAT2026.03 | 14.85 | — | — | 438.5 | — | |
| BiLSTM2026.03 | 15.18 | — | — | 906.71 | — | |
| LSTMYear=2017, Pre-processing steps=Data normalization; RUL target2026.02 | 16.18 | — | — | — | — | |
| TFSCL2026.03 | 16.38 | — | — | 1,180.11 | — | |
| Hybrid modelYear=2021, Pre-processing steps=Feature sel.; norm; piece-wise RUL2026.02 | 16.89 | — | — | — | — | |
| C-Transformer2026.03 | 17.1 | — | — | 939.1 | — | |
| CNN+LSTMYear=2019, Pre-processing steps=Var. threshold; norm; HI2026.02 | 17.12 | — | — | — | — | |
| MATT2025.12 | 18.75 | 15 | 351.56 | — | — | |
| CNNYear=2016, Pre-processing steps=Data normalization; RUL target2026.02 | 19.81 | — | — | — | — | |
| Bi_cLSTM (Ours, 4 blocks)Year=2025, Pre-processing steps=OS norm; z-score by mode; RUL cap=1252026.02 | 21.38 | — | — | — | — | |
| QAQLstatus=proposed2026.06 | 23.44 | — | — | — | — | |
| Quantum VQL2026.06 | 24.66 | — | — | — | — | |
| Quantum DQN2026.06 | 25.34 | — | — | — | — | |
| Quantum AOA2026.06 | 27.49 | — | — | — | — | |
| BACE-RUL2025.03 | 29.36 | — | — | 94,867 | 24.11 | |
| Transformer2026.06 | 29.54 | — | — | — | — | |
| LSTM2025.03 | 29.79 | — | — | 205,346 | 33.46 | |
| Quantum Eigensolver2026.06 | 29.96 | — | — | — | — | |
| PPO2026.06 | 30.05 | — | — | — | — | |
| Deep Q-Network2026.06 | 30.17 | — | — | — | — | |
| Quantum SGD2026.06 | 30.52 | — | — | — | — | |
| DATE2025.03 | 30.96 | — | — | 138,880 | 32.23 | |
| SARSA2026.06 | 31.39 | — | — | — | — | |
| SGD2026.06 | 32.6 | — | — | — | — | |
| Quantum Decision Tree2026.06 | 32.61 | — | — | — | — | |
| Quantum LSTM2026.06 | 32.77 | — | — | — | — | |
| RF2025.03 | 34.79 | — | — | 333,006 | 24.24 | |
| DCNN2025.03 | 34.95 | — | — | 184,826 | 37.51 | |
| SVM2025.03 | 36.14 | — | — | 203,836 | 29.93 | |
| GRU2026.06 | 36.24 | — | — | — | — | |
| Bi-LSTM2026.06 | 39.31 | — | — | — | — | |
| BRR2025.03 | 52.55 | — | — | 1,120,811 | 49.99 | |
| Cox-PH2025.03 | 60.32 | — | — | 1,529,022 | 51.66 | |
| Weibull AFT2025.03 | 73.15 | — | — | 2,887,640 | 62.96 |