Identification of Hoax News in the Using Community TF-RF and C5.0 Tree Decision Algorithm

Authors

  • Enrico Budi Santoso Universitas Jenderal Achmad Yani, Indonesia
  • Yulison Herry Chrisnanto Universitas Jenderal Achmad Yani, Indonesia
  • Gunawan Abdillah Universitas Jenderal Achmad Yani, Indonesia

DOI:

https://doi.org/10.55324/enrichment.v1i6.58

Keywords:

News, Hoax, Decision Tree C5.0, TF.RF, Data Mining

Abstract

News has a great influence on social and political conditions, and news can drive the economy of a country. Identifying hoax news is very important to ensure that the information circulating in society is true and reliable, and helps limit the spread of false information. In the process of reading news spread on social media, people do not know whether it is fact or hoax news because they cannot distinguish whether the news circulating is real news or fake news which if left unchecked can result in the public being misinformed. Therefore, this research process is to create a sistem for identifying hoax news using Decision Tree C5.0, which is an algorithm for the development of the C4.5 algorithm which in a process is almost similar, but the C5.0 algorithm has more value than the C4.5 algorithm which is used for the data mining process with a classification method for 1000 data obtained by web scraping using the keywords "election 2024", "politics" and "checkfaktapilkadamafindo" on the Turnbackhoax.id and Detik.com sites. In this study, what distinguishes it from several previous studies is its existence in several test scenarios, namely classification using feature weighting, which in classification using feature weighting is TF.RF. After testing the confusion matrix on the C5.0 algorithm, it produces accuracy, precision, and recall on each training / test data (70/30) resulting in accuracy 79.33%, precision 80.00%, recall 97.00%, then training / test data (80/20) resulting in accruracy 79.50%, precision 81.00%, recall 95.00%, then training and test data (90/10) resulting in accuracy 72.00%, precision 74.00%, recall 89.00%.

Downloads

Published

2023-09-26