Author(s): Subhadip Chandra; Randrita Sarkar; Sayon Islam; Soham Nandi; Avishto Banerjee; Krishnendu Chatterjee

Sentiment analysis is the methodical recognition, extraction, quantification, and learning of affective states and subjective information using natural language processing, text analysis, computational linguistics, and biometrics. People frequently use Twitter, one of numerous popular social media platforms, to convey their thoughts and opinions about a business, a product, or a service. Analysis of tweet sentiments is particularly useful in detecting if people have a good, negative, or neutral opinion. This study assesses public opinion about an individual, activity, commodity, or organization. The Twitter API is utilised in this article to directly get tweets from Twitter and develop a sentiment categorization for the tweets. This paper has used Twitter data for two separate approaches, viz., Lexicon & Machine Learning. Lexicon based approach further categorized in Corpus-based and Dictionary-based. And various Machine learning-based approaches like Support Vector Machine (SVM), Naïve Bayes, Maximum entropy are used to analyse Twitter data. Neural Network (NN), Decision tree-based sentiment analysis is also covered in this research work, to find out better accuracy of the approaches in the various data range. Graphs and confusion matrices are used to visualise the results of the analysis for positive, negative, and neutral remarks regarding their opinions.