These are the areas of a group project on unstructured vaccination texts that I worked on personally. The project was to analyse a range of texts on vaccination to contribute to the debate. Our group sought to analyse genres of texts to determine the balance of bias within the genre towards the topic.
Areas coverd:
- using CoreNLP for sentiment analysis
- attribution extraction using a list of source indicating predicates (Paretti, 2015)
- feature engineering for stance detection
Pareti, S. (2015). Attribution: A Computational Approach. PhD thesis, University of Edinburgh.