Volume 7, Issue 1, January – 2022 International Journal of Innovative Science and Research Technology ISSN No:-2456-2165 IJISRT22JAN782 www.ijisrt.com 785 Online School Sentiment Analysis in Indonesia on Twitter Using The Naïve Bayes Classifier and Rapid Miner Tools Ahmad Cahyono Adi, Dyan Puji Lestari, Elsa, Fiqrudina Sain Saputri, Yohanes Sabui Faculty of Mathematics and Natural Sciences Tanjungpura University Indonesia Abstract:- The COVID-19 pandemic entered at the beginning of 2020 which hit various countries in the world, including Indonesia, with 4,259,644 contaminated cases (Kawal Covid 19. 2021). The impact of the COVID- 19 pandemic is in the economic, tourism, and education sectors. The most obvious impact due to this pandemic is in the field of education, where every process of teaching and learning activities is limited or even encouraged to study from home. Therefore, teaching and learning activities are carried out online or in a network (online). Educators are starting to look for alternative methods used in online learning because in Indonesia they still use conventional learning or are still in the form of face-to- face learning directly with a classical system.The Naive Bayes method is a classification using probability and statistical methods, namely by predicting future opportunities based on previous experience. The main feature of the Naïve Bayes classification is to get a strong hypothesis from each condition or event. The following is the equation of Bayes' theorem In the rapid miner tools, the data is then retrieved by taking the text and label attributes that have been given during the labeling process. The data is added with a label attribute to facilitate the classification process. After that, the labeled data is then normalized by changing or removing unimportant attributes, in this case, the unimportant data is other than letters. Therefore, a process of deleting non- letter-type data is needed which includes numbers and symbols. Keywords:- Covid-19; Naïve Bayes Algorithm;Sentiment Analysis; Text Mining;Twitter. I. INTRODUCTION The COVID-19 pandemic entered at the beginning of 2020 which hit various countries in the world, including Indonesia, with 4,259,644 contaminated cases [1].The impact of the COVID-19 pandemic is in the economic, tourism, and education sectors. The most obvious impact due to this pandemic is in the field of education, where every process of teaching and learning activities is limited or even encouraged to study from home. Therefore, teaching and learning activities are carried out online or in a network (online). Educators are starting to look for alternative methods used in online learning because in Indonesia they still use conventional learning or are still in the form of face-to-face learning directly with a classical system. In the past year, in Indonesia, there has been a lot of online learning in schools and even campuses in Indonesia. As a simple step, online learning begins by utilizing existing social media, such as Whatsapp, telegram, and youtube. In addition, applications used for the learning process such as Google Classroom, Edmodo, Google Meet, or Zoom [2]. The existence of online schools has attracted a lot of responses from various levels of society, both agree and disagree. Complaints to expressions of pleasure are found on various social media platforms. In addition to complaints and expressions of pleasure, not a few also provide suggestions to improve the education system in online conditions [3]. One of the most widely used social media as a medium for expressing opinions and exchanging ideas is Twitter. Likewise, during the pandemic, Twitter is still one of the social media that has high popularity when compared to other social media. This is evidenced by data from the Ministry of Communication and Information (2021) which states that Indonesia is ranked 6th with 15.7 million Twitter users. The activities of Twitter users who are interested in being active in conversations and dominated by interesting topics lead to interaction so that freedom of expression is built on this social media [4]. Due to a large number of opinions regarding online schools during the Covid-19 pandemic, it is necessary to analyze sentiment on this phenomenon as a consideration so that related parties can consider it in order to overcome the problems that occur. Sentiment analysis is one of the analytical methods from Text Mining that can be used to classify documents in the form of opinion texts based on sentiment [5].Sentiment analysis aims to determine attitudes towards several topics as well as the contextual polarity of the entire document [3]. II. MATERIALS AND METHODS This research method uses the Naïve Bayes method in classifying data to obtain results. The following are the stages of the research described in the image below. Fig. 1: Naïve Bayes Learner