<![CDATA[Indonesian governments have launched PSBB policy to emphasize the growth rate of COVID-19 cases in Indonesia. This caused a variety of responses from the public, one of them on social media Twitter. public opinion contained on social media Twitter, can help the government to know how the public opinion about psbb policy in Indonesia. This study tried to analyze the public's response about PSBB policy on social media Twitter, through sentiment analysts and classified into three classes, namely positive, negative, and neutral. By using Naive Bayes classification method and Improved Gini Index (IGI) feature selection to reduce the number of features in the classification process. The process on sentiment analysis consists of preprocessing, feature selection using the Improved Gini Index (IGI) method, and classification with Naive Bayes. The results of Naive Bayes classification test without feature selection obtained accuracy of 64%, while the results of classification accuracy test with feature selection using six different threshold values obtained the highest accuracy results at the threshold value of 30%, where there are 70% of the total terms removed and obtained accuracy of 68%.]]>
Copyrights © 2021
