Sentiment Analysis on Twitter Dataset using Voting Classifier

Social media platforms like Twitter is now widely used by everyone, where everyone post their opinions regarding products, ongoing trends, politics etc. Sentiment analysis helps to identify the emotional tone of texts. Sentiment analysis of twitter can help to plays an important role to determine and understanding the sentiments, towards various topics, events and trends. This work performs the sentiment analysis on twitter dataset "Sentiment140", a widely used benchmark dataset. Additionally, we explore various techniques for data preprocessing, including text cleaning, tokenization, and feature extraction done by TF-IDF using various techniques such as TF-IDF. Machine Learning models include Multinomial Naive Bayes, Logistic Regression and Gradient Boosting classifier are used for prediction and then these models passed to an Ensemble methodology technique. Voting Classifier is used as an ensemble technique which gives the optimal output than the baseline models.