Data: wtorek, 17.10.2023, godz. 11:00-12:00
Prelegent: Marek Reformat (University of Alberta)
Streszczenie: We present a new method for sentiment analysis that integrates Knowledge Graphs with the content of SenticNet and the Hourglass Model of Emotions. Our goal is to investigate, measure, and classify emotional reactions in reviews. We provide insight into user emotions. The proposed methodology starts by pulling emotional cues from reviews and associating their descriptive terms with various emotional states. Then, we express these using the Hourglass Model of Emotions. All this information is fed into a Knowledge Graph, where every review is linked to its relevant components, descriptors, synonyms, and associated emotions. The findings are supplemented with linguistic expressions of the emotions tied to the identified facets.
Miejsce: B1-7/8 oraz online