Computing Feelings: Danish Approaches to Sentiment Analysis
Location: Dalgas Have, room D1V001 (The workshop is in english)
Time: 13:00-17:00
Sentiment Analysis is new technology for automatically assessing the
opinions and feelings expressed in text. There has been a great deal
of development in recent years, with impressive reports of accuracy
from research labs, and a plethora of commercial offerings. However,
the Danish language has been largely ignored. This meeting brings
together a team of researchers based in Denmark, who have developed a
series of systems in an effort to duplicate or go beyond the current
state of the art in Sentiment Analysis, with a special focus on
Danish.
Program Friday Feb 28, 1pm-5pm
Time | Speaker | Title |
1:00 | Daniel Hardt (CBS) | Introduction |
1:15 | Bernardo Huberman (HP) | Social Media and the Attention Economy |
1:45 | Anders Søgaard (KU) | Representations for sentiment analysis |
2:20 | Break | |
2:30 | Panel Discussion |
Current and Future Business Applications for Language Technology: Bernardo Huberman (HP), Anette Stöckel (Infomedia), Daniel Hardt (CBS) |
3:30 | Break | |
3:40 |
Chris Zimmerman (CBS/MindJumpers),
Yuran Chen (CBS)
|
Marius the Giraffe: Social Media Sentiment |
3:50 | Rene Madsen (Infomedia), Niels Buus Lassen (Evalua) | Ensemble Classifiers |
4:00 | Anders Boje Larsen (Danske Bank) | From simple to more advanced sentiment lexicon approach |
4:10 | Julie Wulff (KU) | |
4:20 |
Claire Joyce (KU),
Kasper Højgaard (KU),
Mattias Sellin (KU),
Mariana Bolano (KU)
|
Sentiment Classification: Domain adaptation for Danish customer reviews |
4:30 |
Jakob Elming (KU),
Dirk Hovy (KU), Barbara Plank (KU)
|
Domain adaptation in Danish Sentiment Analysis |
4:40 | Daniel Hardt (CBS) | Linguistic Structure and Sentiment Analysis |
4:50 | Conclusions |
Everyone is welcome to attend. The workshop should be of interest to people with a technical interest in language technology and also to people who are interested in the practical, business applications of this kind of technology.