Penerapan Algoritma KNN Pada Analisis Sentimen Review Aplikasi Peduli Lindungi

Abstract

Covid-19 made all of Indonesia and even the whole world a country that died from the sars covid-19 virus. The Indonesian government is trying everything for the sake of the community to avoid the outbreak of the sars covid-19 virus. Building an Indonesian government application that is used as a container on the way, the detection of the covid-19 virus was even detected by a vaccination certificate held simultaneously by the Indonesian government. The Pedulilindung application downloaded through the play store provides opinions from various individuals of the community. The author summarizes the Opinin to conduct research on classifying the text of the review, where the amount of data summarized as much as 200 reviews, consisting of 100 positive review data and 100 negative review data, where the sentiment is related to the sentence : good, fast, disappointed, stupid, not worth it. Doing classification with K-Nearest Neighbor method is proven that with this method to get a good accuracy value classification where the accuracy value obtained by 81.72% with AUC 0.856, grouping AUC value 0.856 included into the Good Classification group, so in this case the K-NN method is able to analyze sentiment review PeduliLindungi application.