OPINION MINING AND CLASSIFICATION OF NEW NATIONAL EDUCATION POLICY (2020-2022) USING TWITTER DATA

Abstract

With the growth of technology in recent years, there has also been a rapid increase in the usage of social media sites to exchange information and beliefs. Opining mining, also known as sentiment analysis, is used to ascertain public opinion. It is a technique for natural language processing. Sentiment analysis can be characterised as a technique that utilises natural language processing (NLP) to automate the mining of attitudes, opinions, perspectives, and emotions from text, audio, tweets, and database sources. We gathered data from the microblogging website Twitter regarding the New Education Policy (NEP 2020-2022) in order to have a better understanding of the public mood on a national level. Convenience of social media especially Twitter is that it empowers the swift collection of information about the opinions of the public and individual users on current continuing topics. we employed models to classify and assess the emotion elicited by a compilation of around 22,000 tweets about the topic of New Education Policy (NEP). We carried out our investigation using count vectorizer and TF-IDF and discovered that count vectorizer outperformed TF-IDF. Additionally, we used Naice Bayes, decision trees randon forests, logistic regression, gradient boosting, and Support Vector Machines, and discovered that logistic regression provided the highest assorting accuracy.