Klasifikasi Hoax pada Aplikasi X Mengenai Pemilihan Umum Calon Presiden dengan Metode Adaboost

Abstract

The process of classifying hoax news on the X application related to the presidential election is generally carried out manually by experts. The limitation of valid news sources to verify the truth of a tweet makes readers easily believe information that may not be true. Information technology accelerates the spread of information but can also help readers verify the truth of information through social media, television, and messaging applications. This study aims to determine the accuracy of information and identify the types of hoaxes circulating. Many people still consider news as a hoax if the information is not true. Therefore, this study also aims to inform the public about various types of hoaxes that are widely circulated. This research uses the Adaptive Boosting method with Decision Tree as the base estimator to improve data analysis and processing. The data used comes from tweets on the X application regarding the topic of the presidential and vice-presidential elections in Indonesian. A total of 954 data points were divided into 80% for training data and 20% for testing data. This model achieves an accuracy rate of 91.15% for identifying hoaxes and non-hoaxes, and 83.77% for information classification, with evaluation using cross-validation and evaluation matrics