A study on the sentiments of the Indian Public towards the budget using a machine learning algorithm

  • Aarti Mehta Sharma , Nishigandha Jangnure


Purpose: A lot of hope and aspiration is hinged to economic events like the presentation of the Union Budget by the government as it is expected to bolster economic growth and foster social justice and equity. The highlights of the Indian Budget 2019 were that India is expected to become a US$ 3 trillion economy by the end of this financial year 2019-2020. The paper proposes to ascertain if the public was happier after the budget, lending credence to the second term of the present government


Design/methodology/approach: Various machine learning techniques using Sentiment Analysis have become popular to gauge twitterati’s reactions to major events in the country. This paper proposes a hybrid approach to sentiment analysis with probabilistic topic modelling (using Latent Dirichlet Allocation (LDA)) to examine subjects pertaining to the budget that were most discussed by twitterati during and after the budget. We will further assign sentiments to these topics by using Naïve Bayes algorithm for classification into positive or negative sentiment. The programming software R has been used to extract and analyse tweets. A total of twenty two thousand tweets have been extracted for this study.

 Findings: It was found that people were not pleased with the union budget.

 Originality: This paper explores topic modelling using LDA and shows that tweets can be used to gauge the public’s mood before and after the announcement of an event.