Real-Time Social Media Sentiment Analysis Using Vader and Textblob

This research presents a comprehensive framework of social media data from Facebook and twitter by integrating Natural Language Methods (NLP) with Machine Learning algorithms. Our system employs a rule-based algorithm called VADER for sentiment analysis. To prepare the text data ready for the Machine Learning, it undergoes using stop word removal, tokenization and lemmatization. The system then utilizes TF-IDF, the system then converts the text into numerical features. Using these features, a multinomial Naïve Bayes classifier will be used to predict the sentiment. Additionally, Text Blob is used to determine the subjectivity and polarity of the text. To assess the model's performance, Standard classification metrics, including as accuracy, precision, recall, and confusion matrix are employed.