Classification Hoax News of Covid-19 on Instagram Using K-Nearest Neighbor

Abstract

The world was shocked by the outbreak of a virus known as COVID-19 in Wuhan, China. The spread of the virus is increasingly spreading to all countries in the world, including Indonesia. In Indonesia, information about the COVID-19 virus can be found on various social media, one of which is Instagram. Starting from scratch are found positive cases of COVID-19 in Jakarta. Until now, many news provider's accounts found COVID-19 uploaded on Instagram allows the spread of false news. The spread of fake news has greatly affected the handling of this COVID-19 pandemic. The number of hoax news COVID-19 spread on Instagram makes this study using Instagram as a media data retrieval. Therefore research is needed on classifying hoax news about COVID-19 on news accounts on Instagram. This study uses the K-Nearest Neighbor method to classify COVID-19 news on news accounts on Instagram. This research will classify Instagram based on valid news posts on 30 news accounts on Instagram containing news about COVID-19. The news reference used for categorizing hoaxes is from the @turnbackhoaxid account. In using the K-Nearest Neighbor method, the TF-IDF modeling is applied in conducting assessments and weighting. The research results by using the K-Nearest Neighbor method have a presentation accuracy of success 90%, where the precision value is 83%, recall 100%, and f1-score 91% can solve the problem spread of fake news in Indonesian about COVID-19 on news accounts on Instagram.