Machine Learning Implementation in Sentiment Analysis of Covid-19 Pandemic Handling Policies in Indonesia

Abstract

The behavior of Indonesian society in expressing opinions has shifted towards the use of internet-based social media. The use of social media to voice aspirations also occurred during the emergence of the Covid-19 pandemic in Indonesia, especially in response to the government's policies in handling the outbreak. The purpose of this research is to determine the effectiveness and influence of government policies on the community, one of which is by conducting sentiment analysis on the Twitter social media platform. Previous studies have not specifically analyzed the sentiment of public reactions to government policies in the case of the Covid-19 pandemic or have only discussed the legal aspects of those policies. This research employs a sentiment classification method using machine learning into three categories: positive, neutral, and negative. The analysis is carried out on the occurrence of predefined phrases related to government policies in tweets circulating on social media. The results of this study indicate that public sentiment towards the government's handling of the Covid-19 pandemic tends to be neutral and positive, but the negative sentiment carries a stronger tendency despite its smaller quantity.