Year: 2017
Type: article
Abstract: We present ongoing work that, for the first time, seeks to extract and attribute politicians' quotations from Norwegian Bokmål newspapers.Our methodusing a statistical dependency parser, a few regular expressions and a look-up tablegives modest recall (a best of .570)but very high precision (.978) and attribution accuracy (.987) for a restricted set of speaker names.We suggest that this is already sufficient to support some kinds of important social science research, but also identify ways in which performance could be improved.
Institution: Uni Research (Norway)
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2735353955