IGGSA Workshop on German Sentiment Analysis 2016
Melanie Siegel (Darmstadt University of Applied Sciences, Tutorial)
With the Web 2.0 most consumer products are discussed and evaluated in internet forums. These expressions of opinion contain valuable information for companies: Information about what users think of the product, where they have difficulties in application and how they solve their problems. The sentiment expressions are accessible, but the effort is often too big to regularly read and evaluate them manually. Opinion mining automatically analyzes and classifies sentiment expressions from publicly available sources. For this task, it is not enough to extract keywords from review texts. More intelligent NLP methods are needed to detect negation scope and arguments of opinion words. Therefore, opinion mining is an interesting research topic for language technologists.
Research on opinion mining so far focused mainly on the English language. Some of the approaches can be directly transferred to analyzing German sentiment expressions, but many other cannot.
In the tutorial we examine standard approaches to opinion mining. Further, we look at German sentiment expressions and research approaches to handle them . We will experiment with language data and python programs. We will take into account and introduce existing resources, such as word lists and corpora.
Tutorial Instructor:
Melanie Siegel is professor for information science and semantic technologies at Darmstadt University of Applied Sciences. She holds a Ph.D. in linguistics and a Habilitation in linguistics and computational linguistics from Bielefeld University. Her research interests are grammar development, Japanese linguistics, information extraction, technical documentation, machine translation, and opinion mining.
melanie.siegel@h-da.de