Emoji Prediction using Sentiment Analysis

Emoji prediction is a part of sentiment analysis, and sentiment analysis is a good tool used to classify bodies of text. Reaction to recent events on social media leads to understanding public opinions; hence sentiment analysis can be used to analyze these reactions, thus predicting the public's opinion in advance. The objective of this paper is to examine the dataset from a well-known social media platform and then, based on our analysis, foresee an emoji best suited to the mood and content of the text. To achieve this, we conducted experiments and trained and tested a few AI models on the dataset using supervised learning algorithms. The approach is very straightforward; we took a lot of texts with emojis and trained and tested our models on this data. We find that profound learning calculations (e.g., LSTM and Bi-LSTM) consistently accomplish the best exhibition.