STUDI KOMPARASI ALGORITMA NAIVE BAYES CLASSIFIER DAN SUPPORT VECTOR MACHINE DALAM ANALISIS SENTIMEN TWEET (Studi Kasus : Data Tweet Dengan Kata Kunci ‘psbb’ dan ‘pandemi’)

Starting 2020, the world was shocked by the discovery of a new virus originating from Wuhan, the central city in China. This new virus, including a type of coronavirus, was later named Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-COV2) and caused Corona Virus Disease 2019 (COVID-19). The Indonesian government has implemented various public policies to suppress the spread of the virus, one of which is by imposing Pembatasan Sosial Berskala Besar (PSBB). After a year of the pandemic in Indonesia and the implementation of the PSBB policy, there are many pros and cons regarding the PSBB policy among the public. This study aims to analyze the sentiment or opinion that developed in the community towards the PSBB policy by the government during the COVID-19 pandemic using the Naive Bayes Classifier and Support Vector Machine methods. The data used are 4040 tweets as training data with labels and 6217 tweets as test data without labels. The learning process on the training data resulted in an accuracy value of 62.02% on the Naive Bayes Classifier model and 66.68% on the Support Vector Machine model. The results of the implementation on test data of 6217 tweets using the Naive Bayes Classifier model resulted in a positive sentiment class classification of 40.8% with 2535 data, a neutral sentiment class of 54.8% with 3407 data, and a negative sentiment class of 4.4% with 275 data while using the Support Vector Machine model resulted in a positive sentiment class classification of 38.0% with 2363 data, a neutral sentiment class of 46.8% with 2910 data, and a negative sentiment class of 15.2% with 944 data