How can emotion-ai help understand translator trainees’ technology learning experiences?

The present study examines the effectiveness of Sentiment Analysis, also known as Emotion-AI, in analysing translator trainees’ learning narratives regarding their experiences with translation memory systems (TMs). Students were asked to describe how they learned and whether the experience was pleasant or unpleasant. The narrative texts were then automatically analysed with Sentiment Analysis, and the emotional component was quantified into a Sentiment score which encompasses both the polarity, i.e., positive vs. negative, and the magnitude (in numerical terms) of emotion. The results showed that narratives about pleasant learning experiences had significantly higher scores than those about unpleasant ones, indicating that Sentiment Analysis can be used to identify learners’ emotions while using technology. Our findings suggest that automatic emotion detection tools can be used in combination with human judgments for data triangulation. Keywords: Sentiment Analysis, translation memory, emotions, human-computer interaction