Sentiment Analysis of Twitter Data Using Machine Learning Techniques and Scikit-learn

Sentiment analysis of Twitter data is an area that has experienced significant growth in recent years. The ability to identify sentiment from tweets using machine learning techniques has attracted researchers because of the simple efficiency of machine learning techniques. This paper tackles the use of machine learning algorithms and Scikit-learn in sentiment analysis of Twitter data. To do this, we perform analyses on Twitter datasets made publicly available by NLTK Corpora and create an efficient feature by using a feature extraction technique. We train and test various machine learning classifiers such as MultinomialNB, BernoulliNB, LogisticRegression, SGD classifier, SVC, LinearSVC, and NuSVC. Experimental results demonstrate that BernoulliNB, LogisticRegression, and SGD classifier reached accuracy as high as 75%.