<![CDATA[Current information and communication technology has changed the way information is shared, affecting the way people get and deliver news. The number of digital news that continues to increase every day by several news portals poses a challenge, where news is often related to more than one category. From the existing problems, a study was conducted on the classification of online news titles. This study uses the SVM method with mutual information feature selection to classify online news titles. The dataset used is the news title from detik.com using 6 categories, namely finance, travel, health, auto, food, and sport with the number of data per category being 2000 data. The classification process starts from text preprocessing, term weighting using TF-IDF, then feature selection with mutual information, and finally classification with SVM. The results of the study showed that testing various SVM kernels and mutual information (MI) thresholds with a threshold of 85% provided the highest level of F1-score on the SVM machine with the RBF kernel and a C value = 10, which was 86,15%.]]>
Copyrights © 2025
