Research Article
Mutual-Attention Net: A Deep Attentional Neural Network for Keyphrase Generation
Table 4
Impact of 3 components on predicting the present keyphrase.
| Datasets | Inspec | Krapivin | NUS | SemEval | KP20k |
| Metric | F1@5 | F1@10 | F1@5 | F1@10 | F1@5 | F1@10 | F1@5 | F1@10 | F1@5 | F1@10 | RNN | 0.085 | 0.064 | 0.135 | 0.088 | 0.169 | 0.127 | 0.157 | 0.124 | 0.179 | 0.189 | CopyRNN | 0.278 | 0.342 | 0.311 | 0.266 | 0.334 | 0.326 | 0.293 | 0.304 | 0.333 | 0.262 | PG | 0.281 | 0.342 | 0.313 | 0.267 | 0.333 | 0.330 | 0.291 | 0.303 | 0.335 | 0.264 | MA | 0.085 | 0.065 | 0.138 | 0.089 | 0.170 | 0.126 | 0.159 | 0.125 | 0.182 | 0.194 | MHA | 0.084 | 0.064 | 0.135 | 0.086 | 0.170 | 0.126 | 0.157 | 0.123 | 0.180 | 0.191 | PG + MA | 0.288 | 0.352 | 0.315 | 0.269 | 0.344 | 0.337 | 0.293 | 0.309 | 0.349 | 0.276 | PG + MHA | 0.283 | 0.343 | 0.313 | 0.269 | 0.335 | 0.333 | 0.291 | 0.304 | 0.336 | 0.266 |
|
|