In WeST several researchers have laid their hands on established text mining
techniques to push these techniques forward. Very commonly text does not come alone, but it is actually related to structured data:
- information about the geographical provenance of text,
- relationships to who said it when,
- social networks of people who said it, or
- explicit information about what specific terms imply, e.g. a positive or negative sentiment.
Recent publications and talks now unfold the research that has been going on over the last 2 years. Dive into this work by slides presented at the eDigital conference in Oxford (http://de.slideshare.net/steffenstaab/topic-discovery-in-unstructured-data-the-next-generation).
References are given at the end of the slide deck.