Skip to content Skip to navigation

Qualifying Exam: Fine-Grained Emotion Detection with Emoji

Abstract: 

The usage of emojis is growing day by day everywhere including social media. More than half of the social media text contains emoji these days. In Natural Language Processing, researchers have made great progress in capturing the sentiment and emotion of texts written in different languages. However, these texts do not often come with emojis. Some researchers already showed how emoji carries sentiment in a text or tweet but it still lacks to reveal the emotion of an emoji. In our work, we show how an emoji can help in determining basic human emotions such as anger, fear, joy, and sadness in a tweet. We created a corpus that consists of more than 20 million tweets where every single tweet contains at least an emoji. Later, this corpus was used to produce several frequency-based word embeddings which are indeed being helpful ingredients for determining fine-grained emotions in short texts like tweets.

Speaker: 
Abu Awal Md Shoeb
Location: 
CoRE B (305)
Event Date: 
12/14/2018 - 10:00am
Committee: 
Prof. Gerard de Melo (Chair), Prof. Matthew Stone, Prof. Yongfeng Zhang, and Prof. Ulrich Kremer
Event Type: 
Qualifying Exam
Organization: 
Dept. of Computer Science