Assessing Public Sentiment towards Digital India through Twitter Sentiment Analysis: A Comparative Study
Shaik Suhail Ahmed, Padma Jyothi Uppalapati, Shaik Ayesha, Sajjad Hussain, Kandula Narasimharao, N Silpa
Abstract
As more and more people use various social media platforms like Facebook and Twitter, the amount of data that is available there has been growing daily. These platforms allows users to interact with numerous communities and converse.The data is collected from twitter. Because Twitter data is so heavily unstructured, it is challenging to evaluate. The major goal of this article is to perform sentiment analysis to calculate different levels of polarity of sentiment like positive, negative and neutral and then topic modelling is performed based on different machine learning techniques like logistic regression, Support vector machines, K-Nearest Neighbors, and Decision trees on digital India. Many evaluation criteria, including Precision, Recall, f-score, and Accuracy were used to evaluate the output from these models. Also, this model performs well when extracting texts from Twitter directly for decision trees and logistic regression with accuracy 95.63%.