Embracing Emojis in Sarcasm Detection to Enhance Sentiment Analysis

Emojis are commonly used in social media posts to express sentiment. Most NLP studies clean their text from emojis in the pre-processing stage. This study suggests a framework to analyze sentiment and detect sarcasm by using the Emoji Dictionary (ED) and Sarcasm Detection Approach (SDA) with VADER as a lexicon-based classifier and BERT as a machine learning model to examine the impact of emojis on sentiment classification performance and in detecting sarcasm on three different topic datasets. The study's finding is that keeping emojis, using the ED, and a proposed SDA approach have an effect on increasing classification performance.