Skip to main content
. 2021 Jan 29;16(1):e0245533. doi: 10.1371/journal.pone.0245533

Table 1. Quote extraction evaluation on manually annotated data.

Quotation content Verb Speaker
Precision Recall F1-score accuracy accuracy
Easy match threshold (0.3) 84.6% 82.7% 83.7% 91.8% 86.0%
Hard match threshold (0.8) 77.0% 75.2% 76.1% 93.1% 86.9%