DefinePK

Abstract

With the activities of people in social media giving out data publicly, unstructured data is rising on a day-to-day basis in the form of texts that encompass what people feel and think regarding various issues. Such feelings are of great relevance to researchers, policymakers, and businesspeople keen to know about group behavior. Classical machine learning methods are not always suitable for capturing the specific language of online environments, such as the use of slang, shortcuts, and colloquial language. The latest development in deep learning, utilizing transformers such as RoBERTa, has enhanced the accuracy and context awareness of sentiment classification tasks. The proposed research work would establish an automated environment of sentiment analysis over real-time and bulk text input and with an easily reachable web interface.