Abstract
Due to the advancement and growth of technology, a bulk of unstructured data is generated by different internet users. These users belong to different areas of the world and have different perspectives of thoughts. These users are actively engaged in various social networking platforms (SNPs) like Twitter, Facebook, Instagram, Reedit, etc. Most often, people use these SNPs for amusement or to harm or target others. It is necessary to detect such public emotions in order to protect society. These public emotions heavily rely on the individual's mindset and point of view. The analysis of sentiments is a large area of natural language processing (NLP) that receives a lot of attention for classifying the polarity of the text, figuring out whether it contains subjective information, and detecting what kind of information it is supplying. In this study, We proposed a machine learning-based model for multi-class public emotion detection from Twitter data. The proposed model has been evaluated on the publicly available Kaggle dataset and achieved state-of-the-art performance in terms of precision, recall, and F-measure. The experimental results exhibit that the logistic regression (LR) model has outperformed others.