Table 4 Per class metrics for GPT-3 fine-tuned model
From: Leveraging pre-trained language models for mining microbiome-disease relationships
Class | Precision | Recall | F1-score |
|---|---|---|---|
NA | \(0.570 \pm 0.112\) | \(0.366 \pm 0.108\) | \(0.431 \pm 0.089\) |
Negative | \(0.847 \pm 0.047\) | \(0.881 \pm 0.042\) | \(0.863 \pm 0.041\) |
Positive | \(0.819 \pm 0.045\) | \(0.866 \pm 0.034\) | \(0.841 \pm 0.022\) |
Relate | \(0.820 \pm 0.027\) | \(0.815 \pm 0.022\) | \(0.817 \pm 0.015\) |