Skip to main content

Table 1 Comparison of the inter-annotator agreement and classifier performance over four evaluation measures

From: Stance and influence of Twitter users regarding the Brexit referendum

 

Annotator agreement

Stance classifier

No. of testing examples

6807

37, 048

Alpha

\( 67.7\% \)

\( 45.8\% \)

\(\overline{F_{1}}(-,+)\)

\( 74.6\% \)

\( 60.3\% \)

\({Accuracy}\)

\( 77.2\% \)

\( 59.5\% \)

\({Accuracy}\!\pm \!1\) (\(-,+\))

\( 96.0\% \)

\( 90.5\% \)