数据类型 | 数量 | 方法 | 准确率 | Micro F1 | 精度 | 召回率 |
过程数据 | 1 | SVM | 85.68 | 85.63 | 78.82 | 76.48 |
2 | NB | 86.85 | 86.76 | 76.95 | 89.52 | |
3 | RNN | 91.66 | 91.93 | 90.29 | 93.63 | |
4 | SAE | 91.35 | 89.67 | 90.17 | 89.18 | |
5 | GRU | 92.07 | 92.64 | 93.16 | 92.13 | |
视频数据 | 6 | Deep Video | 52.73 | 47.83 | 52.47 | 43.94 |
7 | C3D | 73.26 | 74.32 | 73.15 | 75.53 | |
8 | P3D | 75.89 | 75.46 | 74.11 | 76.86 | |
9 | S3D | 78.67 | 80.37 | 78.76 | 82.05 | |
10 | TSN | 85.65 | 85.16 | 86.14 | 84.20 | |
11 | Slow Fast | 84.73 | 86.35 | 86.52 | 86.18 | |
12 | TSM | 83.69 | 78.53 | 78.13 | 78.93 | |
13 | TEA | 82.19 | 82.76 | 81.35 | 84.22 | |
14 | VIT | 79.74 | 80.27 | 79.69 | 80.85 | |
15 | Swin Transformer | 80.33 | 81.52 | 80.26 | 82.82 | |
16 | Two Stream Swin Transformer | 92.63 | 93.16 | 92.53 | 93.80 | |
17 | Swinc-Transformer | 90.38 | 91.43 | 91.27 | 91.59 | |
18 | Two Stream Swinc Transformer | 95.26 | 95.37 | 95.68 | 95.06 |