In a study by Roeback, the accuracy of a sentiment analysis system is defined how well it agrees with human judgments

Computer systems have a different understanding of errors than human do the figures are not entirely comparable. Concepts like sarcasm and exaggeration

For practical reasons, using one-dimensional model is very much preferred

Analysis of the effect of sentiment analysis on extracting adverse drug reactions from tweets and forum posts

- due to abundance of text available on social media on ADRs

- using a corpus of posts from the DailyStrength forum and tweets annotated for ADR and indication mentions

- concluded that adding sentiment analysis features can marginally improve the performance of even a state-of-the-art ADR identification method. This improvement can be of use to pharmacovigilance practice, due to the rapidly increasing popularity of social media and health forums.

Korkontzelos I et al. Analysis of the effect of sentiment analysis on extracting adverse drug reactions from tweets and forum posts. J Biomed Inform. 2016 Aug;62:148-58.