Presentation title:
MixedEmotions: Analysis for Social TV, Brand Reputation Management, Call Centres
Presentation description:
We present the MixedEmotions project, which aims at building an open source platform for emotion analysis across big data resources in multiple languages and modalities. The platform is developed around three pilot use cases (Social TV, Brand Reputation Management, Call Centres), each of which have specific requirements for emotion analysis in customer relationship management, i.e., in customer recommendation (Social TV), customer analytics (Brand Reputation Management) and customer monitoring (Call Centres). The MixedEmotions platform builds on a range of language technologies for information extraction, transcription, translation etc., as well as semantic technologies that are based on Linked Data for the purpose of semantic-level information integration. The industry and application partners involved in the project include Expert System and Deutsche Welle (Social TV), Paradigma Tecnologico (Brand Reputation Management), Phonexia (Call Centres) and SindiceTech (semantic information integration and access across pilots). The academic partners provide skills in multimodal emotion analysis (Universitaet Passau), emotion and sentiment modeling (UPM), multilingual text analyis and semantic technologies (Brno University of Technology; National University of Ireland, Galway). The presentation will give an overview of the MixedEmotions approach and present first results in each of the pilots and in platform development.
