Analisis Sentimen Pola Pikir Masyarakat Indonesia Terkait Virus Covid-19 Dalam Media Sosial Twitter Menggunakan Metode Rule Based Leksikon

Abstract

The Covid-19 virus has managed to wreak havoc in various sectors around the world. Not only does it affect a person's physical condition, but also affects the psychological condition of the people's mindset. On the other hand, the Covid-19 pandemic has also had an impact on the high use of social media among the Indonesian people. Individual opinions on various matters are expressed on the web and can be collected into big data for processing. Twitter API is an application created by Twitter to make it easier for other developers to access Twitter web information such as the large amount of data used in this research. Rule-Based Lexicon is a classification method that utilizes rules to distinguish one class from another with positive, negative, and neutral sentiment word classes. This study discusses the results of the sentiment analysis of the mindset of the Indonesian people during the Covid-19 pandemic using the Rule-Based Lexicon classification method with data sourced from Twitter social media as much as 4,068,464 tweet data which obtained an average accuracy value of 81%, a precision value of 93%, 95% recall value, and 83% f1 score. The results of the highest overall mindset sentiment were in July 2021 with 10,011 (23.53%) tweet data classified as positive sentiment on the topic of PPKM in Bangka Belitung Province, 8,629 (19.42%) classified as neutral sentiment on PPKM topic in Gorontalo Province, and 4,216 (8.61%) classified as negative sentiment in Bengkulu Province which leads to a toxic positivity conformity mindset.