KOMPARASI ALGORITMA KLASIFIKASI NAIVE BAYES DAN K-NEAREST NEIGHBORS DALAM ANALISIS SENTIMEN TERHADAP OPINI FILM PADA TWITTER

Abstract

The fact that social media is so unreliable does not prevent Twitter users from using the service. Twitter is one of a few social media platforms that allows users to engage in conversation, share information, or even reveal their true identities, such as when discussing a movie's plot. A tweet or comment about a movie that is posted on Twitter may be viewed as a tool to improve the quality of movie production. To understand this, one can use sentimen analysis to categorize as either negative or positive by comparing the Naive Bayes and k-Nearest Neighbors algorithms to determine which one is the most accurate. The results of the two algorithms' comparative testing reveal that the Nave Bayes algorithm has a higher rata-rata accuracy of 99.63% with an AUC of around 1.000, while the K-NN algorithm has a higher rata-rata accuracy of 99.25% with an AUC of 1.000.