IMPLEMENTASI BIDIRECTIONAL-LSTM DAN WORD2VEC DALAM KLASIFIKASI BERITA PALSU (HOAX) DI INDONESIA

Hidayat, Fadhilah Nur (2026) IMPLEMENTASI BIDIRECTIONAL-LSTM DAN WORD2VEC DALAM KLASIFIKASI BERITA PALSU (HOAX) DI INDONESIA. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21083010082.-cover.pdf Download (1MB)
	Text (Bab 1) 21083010082.-bab1.pdf Download (199kB)
	Text (Bab 2) 21083010082.-bab2.pdf Restricted to Repository staff only until 26 January 2029. Download (645kB)
	Text (Bab 3) 21083010082.-bab3.pdf Restricted to Repository staff only until 26 January 2029. Download (450kB)
	Text (Bab 4) 21083010082.-bab4.pdf Restricted to Repository staff only until 26 January 2029. Download (4MB)
	Text (Bab 5) 21083010082.-bab5.pdf Download (190kB)
	Text (Daftar Pustaka) 21083010082.-daftarpustaka.pdf Download (184kB)
	Text (Lampiran) 21083010082.-lampiran.pdf Restricted to Repository staff only until 26 January 2029. Download (211kB)

Abstract

The spread of fake news (hoax) in Indonesia is increasing with the development of digital technology, potentially causing public unrest and instability. This study aims to build a fake news classification model using a Deep Learning approach with the Bidirectional Long Short-Term Memory (Bi-LSTM) algorithm and Word2Vec. Research data was collected via web scraping from Detik.com for valid news and Turnbackhoax.id for hoaxes, with a total dataset of 2,800 articles used after the balancing process. The pre-processing stage includes text cleaning, normalization, and stopword removal with special handling for negation words. Feature extraction is performed using Word2Vec with Continuous Bag of Words (CBOW) architecture and 300 dimensions, pre-trained on the Indonesian Wikipedia corpus to capture semantic meaning. The Bi-LSTM model is utilized to learn text sequence patterns from two directions (forward and backward). Based on the evaluation of the testing set, the proposed model successfully achieved an Accuracy of 86.43%, with a Precision of 87.23%, Recall of 85.36%, and F1-Score of 86.28% . This research also produced a web-based interface system implementation capable of detecting news validity automatically.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Saputra, Wahyu Syaifullah Jauharis	NIDN0725088601	wahyu.s.j.saputra.if@upnjatim.ac.id
Thesis advisor	Idhom, Mohammad	NIDN0010038305	idhom@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Q Science > QA Mathematics > QA76.87 Neural computers

Divisions:

Faculty of Computer Science > Departemen of Data Science

Depositing User:

Fadhilah Nur Hidayat

Date Deposited:

26 Jan 2026 04:42

Last Modified:

26 Jan 2026 04:42

URI:

https://repository.upnjatim.ac.id/id/eprint/49136