What this software does
HILDA (HIgh-Level Discourse Analyzer) is a discourse parser, it analyzes a text and uncovers the underlying functional relations between its different parts. The text is annotated under a theory of text organization called Rhetorical Structure Theory.
Please note that this software is experimental and probably still contains vicious bugs.
Enter your text
To facilitate the job of the parser, we ask you to kindly indicate each sentence ending by a <s> tag, and each paragraph ending by a <p> tag. For instance, "This is a sentence.<s> Here is another one.<p>" would be a correct input.
See example snippets here or there.
Credits
- Thanks to David du Verle for providing a first version of the parser code, which was developed during his stay at Prendinger Lab. (sponsored by the National Institute of Informatics, Tokyo).
- This software uses the Stanford Parser, libsvm, Penn2Malt, which are the property of their respective authors.