Santoso Heru Agus Primary Author mixed material bibliography Semarang Universitas Negeri Semarang 2020 e English

Artikel Jurnal

hlm : 228-237 Abstract Twitter is a very popular microblog, where users can search for various information, current news, celebrity posts, and hot topics. Indonesia is ranked 5th for the most Twitter users. The large number of users makes Twitter used for the benefit of certain parties with bad goals, such as spreading fake news using fake accounts. Fake accounts are often used by several parties to spread fake news, therefore the spread of fake news must be immediately limited to minimize the negative impact caused by fake news. For this reason, this research is written with the aim of being able to classify fake and genuine Twitter accounts. In this study, using data mining techniques that are closely related to big data in decision making by applying the Naive Bayes method. Naïve Bayes is one of the most widely used classification methods because it has good accuracy and faster computation time. The classification process uses nine parameters, namely based on the Profile Created, Favorite Count, Follower Count, Following Count, Geo Enabled, Follower Rate, Following Rate, Follower Following Ratio, Verified. This study uses 210 datasets of twitter accounts that spread fake news, the result is that Naïve Bayes works very promising  in the classification of fake twitter accounts and in the testing process using 5% of training set produces an accuracy of 80%. Informatika SJI 24077658 Perpustakaan Teknik UPI YAI SJI V7N2 November 2020 SJI2-007 Perpustakaan FT UPI YAI SJI V7N2 November 2020 SJI_V7N2_November_2020.jpg.jpg 58980 2023-02-07 15:41:36 2023-02-07 15:41:36 machine generated