Table 7 The generation quality indexes under different ablation modules of proposed method.
From: Predicting road traffic accident severity from imbalanced data using VAE attention and GCN
Datasets | Indexes | Ablation module | |||
---|---|---|---|---|---|
Attention mechanism | Smooth L1 loss function | Swish activation function | The proposed model | ||
NEWYORK | JSD | 0.2832 | 0.4112 | 0.3892 | 0.1887 |
KLD | 0.2760 | 0.5112 | 0.5133 | 0.1395 | |
WD | 0.4879 | 0.5233 | 0.5342 | 0.2640 | |
CHILI | JSD | 0.2799 | 0.4204 | 0.4144 | 0.1969 |
KLD | 0.2678 | 0.5342 | 0.5311 | 0.2004 | |
WD | 0.5252 | 0.7043 | 0.7209 | 0.2477 | |
BRONX | JSD | 0.2720 | 0.5941 | 0.5911 | 0.1817 |
KLD | 0.2760 | 0.5070 | 0.5912 | 0.1457 | |
WD | 0.4879 | 0.6823 | 0.6123 | 0.2456 |