Sentiment Analysis in Emoji-Rich Text: A Comprehensive Survey of Techniques, Datasets, Challenges and Research Directions 35 28

Abstract

Visual digital communication is advancing substantially, with emoji-rich text, making it quite challenging for the analysis of sentiments from such text. In the paper, a comprehensive overview is provided for the methodologies for sentiment analysis in emoji-rich text. In this study, the landscape of available datasets curated for this text is reviewed. Their relevance, limitations, and linguistic diversity is also studied. Furthermore, a comprehensive review of the preprocessing techniques and the machine learning models utilized is also presented. A detailed study is conducted to highlight the specialized processing techniques used for the interpretation and handling of emojis, such as emoji embeddings and ranking systems. Further, research gaps are highlighted, including a multilingual and multimodal approach, and non-availability of high quality datasets. The paper concludes with a discussion on future work advocating for context-sensitive models and the use of generative AI for the development of emoji-aware pre-trained models. This work will serve as a comprehensive reference point for researchers planning to advance sentiment analysis of the emoji-driven digital world.