Improving NER models using gazetteers. The result is only for 3 Tags: PER, ORG, & LOC. Models trained for 50 epochs. Result is an average over 5 runs.
| Method . | amh . | hau . | ibo . | kin . | lug . | luo . | pcm . | swa . | wol . | yor . | avg . |
|---|---|---|---|---|---|---|---|---|---|---|---|
| CNN-BiLSTM-CRF | 50.31 | 84.64 | 81.25 | 60.32 | 75.66 | 68.93 | 62.60 | 77.83 | 61.84 | 66.48 | 68.99 |
| + Gazetteers | 49.51 | 85.02 | 80.40 | 64.54 | 73.85 | 65.44 | 66.54 | 80.16 | 62.44 | 65.49 | 69.34 |
| Method . | amh . | hau . | ibo . | kin . | lug . | luo . | pcm . | swa . | wol . | yor . | avg . |
|---|---|---|---|---|---|---|---|---|---|---|---|
| CNN-BiLSTM-CRF | 50.31 | 84.64 | 81.25 | 60.32 | 75.66 | 68.93 | 62.60 | 77.83 | 61.84 | 66.48 | 68.99 |
| + Gazetteers | 49.51 | 85.02 | 80.40 | 64.54 | 73.85 | 65.44 | 66.54 | 80.16 | 62.44 | 65.49 | 69.34 |