Using Textual Emotion Extraction in Context-Aware Computing

Abstract

In 2016, the number of global smartphone users will surpass 2 billion. The common owner uses about 27 apps monthly. On average, users of SwiftKey, an alternative Android software keyboard, type approximately 1800 characters a day. Still, all of the user-generated data of these apps is, for the most part, unused by the owner itself. To change this, we conducted research in Context-Aware Computing, Natural Language Processing and Affective Computing. The goal was to create an environment for recording this non-used contextual data without losing its historical context and to create an algorithm that is able to extract emotions from text. Therefore, we are introducing Emotext, a textual emotion extraction algorithm that uses conceptnet5’s realworld knowledge for word-interpretation, as well as Cofra, a framework for recording contextual data with time-based versioning.