Sentiment Analysis Regarding the Impact of Covid-19 on Education in Indonesia with The Naïve Bayes Classifier

In December 2019, the new coronavirus SARS-CoV-2 unexpectedly produced the COVID-19 pandemic in China. The World Health Organization reports millions of verified cases and more than a hundred thousand confirmed deaths globally. As a result, during various outbreak-related occurrences, these social media platforms are exposed to and display a variety of perspectives, ideas, and feelings. Lately, there have been many tweets from the public regarding the dire conditions that have occurred in Indonesia since the COVID-19 outbreak, especially related to its impact on the education sector in Indonesia, which until now still often causes pros and cons in most surrounding communities. One of the effective classification methods for sentiment analysis methods is Nave Bayes. Naïve Bayes is applied by determining the appearance of sentiment contained in tweets. Before applying the Naïve Bayes method, data preprocessing and application of the TF-IDF method were carried out. From the research that has been done, the accuracy rate of the Naive Bayes Classifier algorithm is 0.696 with a procedure without the stemming stage and 0.705 with the stemming stage present. The presence or absence of the stemming process significantly impacts the classification's outcome; in this study, the classification's final value was slightly higher when the stemming step was used than when it was not