Model
This is a BioBERT based model trained on a set of manually annotated texts with causation labels, tasked with classifying a sentence into different levels of strength of causation.
This rating-pubmed version is tuned on the dataset provided in a published article Yu et al. (2019) Detecting Causal Language Use in Science Findings.
- Downloads last month
- 9
Model tree for kelingwang/bert-causation-rating-pubmed
Base model
dmis-lab/biobert-base-cased-v1.2Evaluation results
- off by 1 accuracy on pubmed_textdataself-reported83.562
- mean squared error for ordinal data on pubmed_textdataself-reported0.811
- weighted F1 score on pubmed_textdataself-reported0.821
- Kendall's tau coefficient on pubmed_textdataself-reported0.793