<![CDATA[The existence of a swab-PCR service must also consider various inputs, one of which by looking at the public (community) response the swab-PCR service. Through this explanation, this study uses tweet data containing responses, resulting in classification and predictions that be used a benchmark for evaluating swab-PCR services. With this in mind, study will conduct sentiment analysis by classifying tweet data words into positive and negative words. The approach used is the Support Vector Machine (SVM) algorithm. Going through data mining process with RapidMiner, labeling the mining data, followed by text-preprocessing, then data divided into training data (80%) and testing (20%), tf-idf weighting for each word in the data in the form numeric, classification process with SVM, evaluation with confusion matrix and validation with cross validation, visualization with wordcloud and pie charts. So that confusion matrix generated validation with average score of 66% which is tested with an accuracy of 76%, precision of 75%, and recall of 81%. Then based wordcloud obtained four positive words are gratis, sehat, mandiri, dan positif, and four negative words are kecewa, mahal, tolak, dan antri. The percentage of positive sentiment of 54.4% and the percentage negative sentiment 45.6% are visualized through the pie chart.]]>
Copyrights © 2022
