COMPARISON OF K-NEAREST NEIGHBOR AND NAIVE BAYES METHODS FOR CLASSIFICATION OF NEWS CONTENT

Abstract

With the development of technology, news reading via the internet or digital tends to increase. In addition, there are about 300 to 400 news articles in one month and many categories of news articles in a web portal. It makes the editor's performance more and more because an editor must be able to edit articles from various channels and at the same time have to categorize articles one by one manually into several specified categories. This study aims to compare the K-Nearest Neighbor (KNN) and Naive Bayes methods to classify news content in order to obtain the best method. The data used in this study are news articles from the web portal kaltimtoday.co from January 2022 to March 2022. Therefore 576 data are obtained. The results showed that the application of the KNN and Naive Bayes methods could be used to classify news content. The KNN method is able to produce a higher accuracy value than Naïve Bayes, reaching 86% and 51% with test data of 100 news articles.