Prediction of Political Biasness of Statements

Sentiment analysis, commonly referred to as opinion mining, is a branch of natural language processing (NLP)that deals with identifying and extracting subjective information from text, such as opinions, attitudes, and emotions. In this study, the main aim to develop a machine learning-based model to predict the political biasness of tweets. Tweets related to different political ideologies like leftist, rightist and neutral are extracted utilizing the snscrape module on basis of hashtags and usernames. Text preprocessing is done using the NLTK library involving steps like stop words removal, lemmatization and stemming. Next, the text in the tweets is transformed into usable vector forms using the count vectorizer technique. Then train test splitting is done. Then, six classification techniques namely - Multinomial-NB, Logistic-Regression, Ridge Classifier, Random Forest classifier, AdaBoostClassifier and Linear SVM are applied on the dataset. The best accuracy obtained is 0.807 with Logistic Regression.