Sentiment prediction from indian political news content

Abstract

Twitter is a widespread micro-blogging, social media platform used for political campaigning, and in India, it is among the largest online platforms used for sharing political information and expressing ideas, and opinions about politics which gain popularity when liked by a large number of users. In today social media users create a huge amount of unstructured text in the form of messages, posts, and blogs. The aim of this paper is to predict the subject of political news and sentiment of that news by using a ml model which is trained by the 2019 twitter data on indian politics. In this research paper, we use 2019 Twitter data and find the polarity of every tweet. We collect 28,000 political headlines from 4 news websites .First we collect dataset from kaggle then we preprocess the tweet then using some machine learning model we predict the subject of tweet and sentiment of it .Lastly we use this model to predict the subject and sentiment of politics news. The newspaper data was collected using Beautiful Soup python library which scrap the news from various web site. The sentiment analysis was performed using a dictionary-based lexicon approach in conjecture with a “bag-of-word” and “TF-IDF” and vector space model.