JOURNAL

Layanan journal yang disediakan oleh Perpustakaan Universitas Gunadarma

Deteksi Komentar Spam Bahasa Indonesia Pada Instagram Menggunakan Naive Bayes

Judul Artikel:Deteksi Komentar Spam Bahasa Indonesia Pada Instagram Menggunakan Naive Bayes
Judul Terbitan:Jurnal Teknik Informatika (Ultimatics)
ISSN:20854552
Bahasa:IND
Tempat Terbit:Tangerang, Banten
Tahun:0000
Volume:Vol. 9 Issue 1 0000
Penerbit:Universitas Multimedia Nusantara (UMN)
Frekuensi Penerbitan:-
Penulis:Antonius Rachmat c., Yuan Lukito
Abstraksi:Abstract- Instagram is the most famous pictures and videos media sharing based on the web & mobile application. Instagram users can have picture posts that can be commented by their followers. Indonesian public figures such as actors, actresses, musicians use Instagram to promote their activities to their followers. Unfortunately, there are a lot of spam comments in Instagram that need special attention and have to be removed. This research grabs Instagram comments and builds tire dataset from Indonesian public figures who have more than one million followers. By using preprocessing (tokenization, stop words removal, and stemming), TF-IDF weighting, and supervised learning, Naive Bayes method is used to detect spam comments in Indonesian. Naive Bayes produces 74,31% accuracy rate on unbalanced datasets and 77,25% accuracy rate on balanced datasets. This result shows tlrat Nuive Bayes can be used to build all automatic Indonesian spam comments detector on Instagram with high accuracy rate. The novelty of this research is that Naive Bayes can be used to detect spam comment on our Indonesian Instagram comments dataset.
Kata Kunci:Instagram; Naive Bayes; Indonesian spam comments; spam comments detection
Lokasi:P50
Terakreditasi:belum