ABSTRACT
Another crisis has emerged in the shape of widespread anxiety and panic, fueled by imprecise and frequently incorrect information, during the Coronavirus pandemic. As a result, there is a critical need to address and better comprehend COVID-19's informational crisis, as well as evaluate public mood, in order to adopt effective communications and policy decisions. This study aims to classify the results of PIKOBAR's review sentiments, PIKOBAR is a Center for Information and Coordination of Diseases and Disasters in West Java. A total of 371 reviews were taken, each of which was labeled positive, negative or neutral. The data first goes through a pre-processing process before conducting a sentiment review analysis using the Naive Bayes Classifier and Support Vector Machine processes. The results from 80% testing data and 20% training data obtained the Naive Bayes accuracy rate of 75.67% and the Support Vector Machine was 71.62%. Furthermore, in the text association process, information was obtained that the PIKOBAR application users mostly talked about words 'Jabar' for positive class and the word 'aplikasi' for negative class and the word 'data' for neutral class. © 2022 IEEE.