Table 1 Comparison of the sentiment polarity mark-ups derived from two methods, rule-based and trained BERT, before and after the keyword update.

From: New drugs and stock market: a machine learning framework for predicting pharma market reaction to clinical trial announcements

 

# Divergences

# Coinciding positives

# Coinciding negatives

# Coinciding neutrals

With the initial keywords

207

1447

445

3337

With the updated keywords

66

1562

765

304

  1. The number of divergences represents the number of announcements labeled differently by these methods. Whereas, for example, the number of coinciding positives reflects the number of announcements that are positive from the perspective of both methods.