Activity Recognition on PAMAP2
99.04AccuracyWeakly Self-Supervised Approach with 10% Labels
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Weakly Self-Supervised Approach with 10% LabelsSupervision Level=Weakly Self-Supervised, Label Percentage=10%2025.12 | 99.04 | 97.82 | — | |
| Weakly Supervised Multi-Task ApproachSupervision Level=Weakly Supervised, Task Configuration=Multi-Task2025.12 | 98.93 | 97.66 | — | |
| Weakly Supervised Single-Task ApproachSupervision Level=Weakly Supervised, Task Configuration=Single-Task2025.12 | 98.56 | 97.36 | — | |
| Weakly Self-Supervised Approach with 5% LabelsSupervision Level=Weakly Self-Supervised, Label Percentage=5%2025.12 | 97.55 | 96.31 | — | |
| Fully Supervised TCNSupervision Level=Fully Supervised2025.12 | 95.27 | 93.82 | — | |
| Weakly Self-Supervised Approach with 1% LabelsSupervision Level=Weakly Self-Supervised, Label Percentage=1%2025.12 | 91.28 | 89.67 | — | |
| Contrastive GNN2025.12 | 86.86 | 86.36 | — | |
| Efficient Deep Clustering2025.12 | 86.3 | 80.1 | — | |
| Self-Supervised ApproachSupervision Level=Self-Supervised2025.12 | 85.43 | 82.59 | — | |
| Self-Attention2025.12 | 83 | 81 | — | |
| ResGCNN2025.12 | 82.18 | 81.97 | — | |
| DeepConvLSTMSupervision Level=Fully Supervised2025.12 | 80.53 | 77.92 | — | |
| AutoencoderSupervision Level=Self-Supervised2025.12 | 77.06 | 73.98 | — | |
| UniMTSEvaluation Protocol=Zero-Shot2024.10 | 47.2 | 43.6 | 63.2 | |
| UniMTS w/o text augEvaluation Protocol=Zero-Shot2024.10 | 38.4 | 33.9 | 48.8 | |
| UniMTS w/o rot augEvaluation Protocol=Zero-Shot2024.10 | 32.5 | 28.3 | 48.8 | |
| UniMTS w/o graphEvaluation Protocol=Zero-Shot2024.10 | 30 | 23.6 | 41.8 | |
| ImageBindEvaluation Protocol=Zero-Shot2024.10 | 14.6 | 8.1 | 15.6 | |
| HARGPTEvaluation Protocol=Zero-Shot2024.10 | 11.1 | 2.1 | 23 | |
| LLaVAEvaluation Protocol=Zero-Shot2024.10 | 10.3 | 1.6 | 18.7 | |
| IMUGPTEvaluation Protocol=Zero-Shot2024.10 | 8.9 | 1.5 | 19.3 | |
| IMU2CLIPEvaluation Protocol=Zero-Shot2024.10 | 1.9 | 1.1 | 9.8 |