Lexicon and Heuristics Based Approach for Identification of Emotion in Text
Junaid Akram, Arsalan Tahir · 2018
Recognition of emotion from text is emerging as new field of research which can add another dimension to the analysis of textual data. People directly or indirectly express their emotion through facial expressions, speech or written content. People are putting a lot of textual content on social media and microblogging platforms. This data can be very useful in discovering different aspects including emotions. Recognition of emotion is a very complex task. We propose a technique which uses word lexicons, emoticons, negations and intensity modifiers. Our approach follows Ekman's emotion model. Word lexicons are generated using WordNet from a set of seed words collected manually. Twitter is used to collect emoticons. Negations and intensity modifiers are collected manually from literature. Several heuristic rules are designed to support our keyword spotting technique. Experiments on data obtained from Twitter shows high precision and recall for all emotional classes.