Deteksi Berita Hoax Dan Disinformasi Covid-19 Menggunakan Metode Naïve Bayes Berbasis Particle Swarm Optimization

Abimanyu, Putra Prassiesa (2023) Deteksi Berita Hoax Dan Disinformasi Covid-19 Menggunakan Metode Naïve Bayes Berbasis Particle Swarm Optimization. Undergraduate thesis, Politeknik Negeri Jember.

	Text (Abstract) 1. E41192178 Putra Prassiesa Abimanyu - Abstrak.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (50kB)
	Text (Bab 1 Pendahuluan) 2. E41192178 Putra Prassiesa Abimanyu - Bab 1.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (133kB)
	Text (Laporan Lengkap) 3. E41192178 Putra Prassiesa Abimanyu - Full Text.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text (Daftar Pustaka) 4. E41192178 Putra Prassiesa Abimanyu - Daftar Pustaka.pdf - Submitted Version Available under License Creative Commons Attribution Share Alike. Download (183kB)

Abstract

This research develops a hoax and disinformation detection system using the Naive Bayes method based on Particle Swarm Optimization (PSO). Researchers used news data as much as 3000 with the distribution of training and test data 80:20. The N-gram variation of TF-IDF is used with unigrams, bigrams, and trigrams. The Naive Bayes method is also varied with three smoothing techniques: Laplacian, Lidstone, and without smoothing. Feature selection is optimized using PSO. The experimental results show that the PSO-based Naive Bayes method can improve detection performance with an accuracy of 72.38%, a precision of 71.75%, a recall of 72.38%, and an F1-score of 71.9%. Using the TF-IDF unigram gives better results. The Lidstone method in smoothing shows the best performance, and optimizing feature selection with PSO can improve detection accuracy.

Actions (login required)

View Item