1. The Application
The GATE Generic Sentiment Analysis application is a base tool for a rule-based sentiment analysis application which is designed to be adapted by the user to a particular domain or task. For a discussion of why this adaptation process is critical, see . The application cannot be run as standalone, but must be either paired with an entity or term extraction tool such as ANNIE, TwitIE, or TermRaider, or run on a corpus already annotated with linguistic components such as Tokens, POS Tags and entities or terms.
2. Try a Demo
Try this simple demo of an adaptation of the tool for environmental tweets.
3. Try it on the Cloud
Try a version of the application for tweets as a GATE Cloud service.
4. Run the application in GATE
The easiest way to run the application is to use GATE 8.6-SNAPSHOT (or above) as the application is provided within the Sentiment plugin.
If you need to use an older version of GATE then you can download an appropriate version of the application:
5. More information
Various domain-specific sentiment analysis tools have been developed in GATE over the years, such as the environmental application ClimaTerm, and the Brexit analyser. For each application, the main generic sentiment analysis components are re-used and re-purposed. For example, the kinds of things we want to find opinions about may differ (political topics vs environmental terms), the ways in which opinions are expressed may differ, different kinds of expressions may be used, and the opinionated words may have different polarity or meaning depending on the domain and context. The generic sentiment analysis tool can be used as it is, but its performance will be enhanced when some adaptation is done. Instructions for how to adapt the various components, and more information about the composition of the pipeline, are available here.
There is currently no specific reference to be cited that describes the application. However, the following references refer to elements of the work. See the GATE publications page for more relevant references.
 D. Maynard and K. Bontcheva. Challenges of Evaluating Sentiment Analysis Tools on Social Media. In Proc. of Language Resources and Evaluation Conference (LREC), May 2016, Portoroz, Slovenia. PDF
 Diana Maynard and Jonathon Hare. Entity-based Opinion Mining from Text and Multimedia. In "Advances in Social Media Analysis", Mohamed Gaber, Nirmalie Wiratunga, Ayse Goker, and Mihaela Cocea (eds.) 2015, Springer. Springer link.
 D. Maynard and K. Bontcheva. Understanding climate change tweets: an open source toolkit for social media analysis. In Proc. of EnviroInfo 2015, Copenhagen, Sep. 2015. PDF
 Diana Maynard, Gerhard Gossen, Marco Fisichella, Adam Funk. Should I care about your opinion? Detection of opinion interestingness and dynamics in social media. Journal of Future Internet, Special Issue on Archiving Community Memories, 2014. PDF
 Diana Maynard and Mark A. Greenwood. Who cares about sarcastic tweets? Investigating the impact of sarcasm on sentiment analysis. Proc. of LREC 2014, Reykjavik, Iceland, May 2014. PDF
7. License information
License files are included that indicate licensing status of the components. The plugins directory in the pipeline download contains all the plugins required to run the application; note that some of these may have a different licensing status to the main application.
Research supported by the European Union under grant agreement No. 610829 Decarbonet and No. 654024 SoBigData.