An Emotional Topic Recommendation Chat Assistant

Abstract

Mobile technology advances such as online chat or messaging apps were created to make communication easier and more convenient. However, the absence of tone and non-verbal cue, and lack of warmth, etc. in online communication results in shallow interactions, less demand for communications, and could cause negative effects on social relations. Despite a growing body of research about the design and use of recommender systems in online chat scenarios, existing work has paid a little attention to help interactions and social relations between users of messaging Apps. In this paper, we present a chat assistant trying to address this problem by recommending positive and negative emotional topics to users during online chat. As the result, users can take the recommended topics into account during the conversation to have more positive emotional connections and social relations. Our preliminary experiment shows an acceptable accuracy of the recommendation system design in different conversation scenarios.