Analisis Klaster Judul Berita di Website Suarapublik.Id Menggunakan Metode K-Means

Abstract

The rapid development of online news media has led to a significant increase in the amount of information published daily, requiring analytical methods capable of systematically organizing information. Suarapublik.id, as an online news platform operating in the field of information and communication, publishes diverse news headlines covering various topics, which may complicate the identification of content patterns. This study aims to cluster news headlines on the suarapublik.id website using the K-Means Clustering method based on text mining techniques. The literature review of this research is grounded in the concepts of text mining, text preprocessing, TF-IDF weighting, and the K-Means algorithm as a text clustering method. This research employs a quantitative approach with 100 news headlines collected manually and processed using the Google Colab platform. Data analysis was conducted through text preprocessing, TF-IDF weighting, determination of the optimal number of clusters using the Elbow Method, and K-Means clustering. The results show that seven distinct clusters were formed, each representing different news themes, reflecting the content patterns and topic tendencies on suarapublik.id. This study demonstrates that the K-Means method is effective in automatically and systematically clustering news headlines.