Table 2 F1 score
From: Enhancing argumentation component classification using contextual language model
Language Model | Class | Average | |||
|---|---|---|---|---|---|
Major Claim | Claim | Premise | None | ||
Bert Base Cased (Seq. 80) | 0.54 | 0.50 | 0.80 | 1.00 | 0.71 |
Bert Base Cased (Seq. 256) | 0.47 | 0.47 | 0.82 | 1.00 | 0.69 |
Bert Base Uncased (Seq. 80) | 0.53 | 0.49 | 0.82 | 1.00 | 0.71 |
Bert Base Cased without Fine Tuning | 0.00 | 0.31 | 0.00 | 0.00 | 0.08 |
Elmo | 0.47 | 0.23 | 0.79 | 0.98 | 0.62 |
Distilbert-base-cased (Seq. 80) | 0.44 | 0.44 | 0.79 | 1.00 | 0.67 |
Distilbert-base-uncased (Seq. 80) | 0.44 | 0.43 | 0.80 | 1.00 | 0.67 |
Word2Vec | 0.10 | 0.31 | 0.66 | 0.44 | 0.38 |
GloVe | 0.21 | 0.21 | 0.67 | 0.43 | 0.38 |
FastText | 0.10 | 0.00 | 0.72 | 0.44 | 0.32 |