Understanding Challenges Presented Using Emojis as a Form of Augmented Communication

Emojis are frequently used in social media and textual conversations. They are significant means of communication to visually help express emotions and describe objects. Previous studies have shown positive impacts of emojis used in human relations, memorization tasks and engagement with web content. Unicode version 6 includes 2923 emojis, which makes it difficult to use them effectively without a recommender system. We formulate recommending emojis as a complex prediction problem based on its diverse usage as a word and as a sentiment marker. People have individual usage and representations of emojis across different platforms as well as different interpretations based on the device. In order to increase the accuracy of the recommending process, we propose using a recommender system that applies personalization to suggest various emojis. This paper describes several baseline models we explored and the Long Short-Term Memory (LSTM) recurrent neural networks we trained to test the feasibility of extracting knowledge from emoji datasets to predict emoji usage.