Kurzfassung

An AI-software for analyzing Social Media Data with reference to emotional states related to automotive brands was developed as a result of a cooperation between Fraunhofer IAIS and Prof. N. Krautkrämer-Merkt.

The EmotionsRadar concept was based on a dissertation on emotional customer satisfaction (Ph.D thesis 2007, see publications below). The categorization of emotions according to the Emotions-Mix model developed in this research in combination with Fraunhofer's KI Technology lead to...An AI-software for analyzing Social Media Data with reference to emotional states related to automotive brands was developed as a result of a cooperation between Fraunhofer IAIS and Prof. N. Krautkrämer-Merkt.

The EmotionsRadar concept was based on a dissertation on emotional customer satisfaction (Ph.D thesis 2007, see publications below). The categorization of emotions according to the Emotions-Mix model developed in this research in combination with Fraunhofer's KI Technology lead to models that were able to detect emotions in social media text data.

After many years of research, the first algorithm AI-prototype was programmed from 2015-2017 with Conditional random fields probabilistic models for segmenting and labeling sequence data by Fraunhofer IAIS. The necessary test data and its classification was delivered by EmotionsRadar GmbH.

In the end of 2017, after nearly ten years of research, the algorithm was able to detect emotions in text and to differentiate between ten types of emotions. The prototype was tested and further developed from 2017 to 2019 on 10 Mio. customer postings from fora dealing with automotive brands.

In 2019 a Big Data application for social media monitoring was designed and results presented to potential customers to get their feedback on the analysis, e.g. BMW. As the data protection regulations finally got stricter and stricter from 2008-2020, further steps were not undertaken, despite positive feedback from the industry. However, a proof-of-concept herewith was made.

Concerning Fraunhofer, the research was supported by Theseus (EU project).
» weiterlesen» einklappen