- Lecturer: OHMAN, Emily
*Lecturer Information: https://w-rdb.waseda.jp/html/100002999_en.html - Title: Sentiment analysis for interdisciplinary projects
- Day and Time: November 17,2022 6:00pm –
- Abstract:
Sentiment analysis is the automated process of extracting affect and emotions from data. There are several challenges facing sentiment analysis, including when used with machine learning. For supervised machine learning, labeled or annotated data is necessary, however, humans only agree upon an emotion label around 70-80% of the time so how can we teach a computer something humans are not certain of or agree upon? This makes it not only expensive to source annotations, but even more difficult to source reliable annotations in sufficient quantities. I will discuss the general process of computational affect studies followed by a presentation on some recent interdisciplinary projects of mine that have explored affective narratives in different domains, including my Kakenhi project: “Negative Emotions in Literature: A Computational Approach to Tone and Mood.”
