Sentiment Analysis on Citizenship Amendment Act of India 2019 Using Twitter Data

Abstract

For the perspective of the latest happening news or some events occurring worldwide, social media is widely used and the reaction given by the people’s opinion is in the form of raw natural data in many languages and environments. All those written views have some unbalanced statements, i.e., some sensitive information or some slang words and uneven words. This makes opinion mining and making strategic decision useful in the future market. The structured and unbalanced data, Natural Language Processing (NLP) and Data Mining techniques are used for sentiment analysis. In the developed method, the study focuses on Twitter data on Citizenship Amendment Act of India, 2019 to detect the sentiment of the views from people all over the world using machine learning techniques. Many people had given their opinions and views about this new rule for CAA throughout that time. By purifying and analyzing the data using NLP techniques, VADER (Valence Aware Dictionary and sEntiment Reasoner) sentiment polarity is calculated. The dataset is normalized to be used by machine learning algorithms and prepared using natural language techniques such as Word Tokenization, Stemming and lemmatization, and Part of Speech (POS) Tagging. All that input variables are converted into vectors by Term Frequency-Inverse Document Frequency (TF-IDF). This method is implemented in a python programming language. The evaluation parameters such as accuracy, precision, recall, and F1-score were obtained for Naïve Bayes, SVM (support vector machine), K-Nearest Neighbor, Neural Network, Logistic Regression, Random Forest and LSTM (Long-short Term Memory) based RNN (Recurrent Neural Network). Finally, the results are compared. A One-way Analysis of Variance (ANOVA) test was performed on the mean values of performance metrics on all the methods.