Skip to main content
. 2023 Mar 29;10:175. doi: 10.1038/s41597-023-02089-z

Table 3.

Compare the F1 scores of entity recognition in various models.

Entity(freq. in test set) MaxEnt BiLSTM-CRF BiGRU-CRF BERT-CRF BERT-BiLSTM-CRF
MATERIAL(92) 43.37 49.56 50.40 57.58 57.96
METHOD(97) 37.97 46.35 47.88 56.45 57.41
PRODUCT(94) 68.25 81.88 82.16 89.97 90.86
FARADAIC EFFICIENCY(62) 83.68 87.56 87.98 92.12 92.68
Macro-avg F1 49.23 64.44 65.58 69.47 70.12
Micro-avg F1 68.03 78.03 80.69 81.56 81.95