Crowdsourcing and Validating Event-focused Emotion Corpora for German and English

摘要: Sentiment analysis has a range of corpora available across multiple languages. For emotion analysis, the situation is more limited, which hinders potential research on cross-lingual modeling and development predictive models for other In this paper, we fill gap German by constructing deISEAR, corpus designed in analogy to well-established English ISEAR dataset. Motivated Scherer's appraisal theory, implement crowdsourcing experiment consists two steps. step 1, participants create descriptions emotional events given emotion. 2, five annotators assess expressed texts. We show that transferring an classification model from original crowdsourced deISEAR via machine translation does not, average, cause performance drop.