Analisis Sentimen Publik terhadap Kebijakan Pemberlakuan Pembatasan Kegiatan Masyarakat skala Mikro Menggunakan Algoritma Support Vector Machine (Studi Kasus: Twitter)

Abstract

All countries in the world are experiencing difficult times due to the Covid-19 pandemic. Crises are emerging in the health and economic sectors, including Indonesia. The president requested that this crisis be addressed immediately and should be resolved simultaneously. Various efforts have been made by the government such as conducting PPKM Mikro. PPKM Mikro policy raises the pros-cons of society because of its impact on the economy and is ineffective in reducing daily cases of Covid-19. Public responses and opinions are conveyed through various media, more than 63.6% of total social media users in Indonesia have a Twitter account. Based on this background, this study hopes to use data from Twitter to understand the response and perception of Indonesian people to PPKM Mikro policy, from the prediction results, which can be a benchmark or evaluation material for the government. This study conducted a sentiment analysis by dividing people's responses into positive and negative sentiments using the Term Frequency Inverse Document Frequency algorithm for weighting and the Support Vector Machine method for classification. The results of the research that has been done showed the validation results using k-fold cross-validation on the support vector machine has an average cross-validation score of 96.42%. The test results showed an accuracy of 97.13%, precision of 97.58%, and recall of 99.15%. The overall percentage of sentiment analysis is for negative 12.8% and for positive 87.2%.