IMPLEMENTATION OF A BROWSER EXTENSION FOR BILINGUAL CYBERBULLYING DETECTION ON SOCIAL MEDIA USING BILSTM AND LIME EXPLANATION

Syahputra, Mochammad Daffa Faiq Husin (2026) IMPLEMENTATION OF A BROWSER EXTENSION FOR BILINGUAL CYBERBULLYING DETECTION ON SOCIAL MEDIA USING BILSTM AND LIME EXPLANATION. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 22081010161.-cover.pdf Download (4MB)
	Text (Bab 1) 22081010161.-bab1.pdf Download (133kB)
	Text (Bab 2) 22081010161.-bab2.pdf Restricted to Repository staff only until 21 July 2029. Download (1MB)
	Text (Bab 3) 22081010161.-bab3.pdf Restricted to Repository staff only until 21 July 2029. Download (933kB)
	Text (Bab 4) 22081010161.-bab4.pdf Restricted to Repository staff only until 21 July 2029. Download (8MB)
	Text (Bab 5) 22081010161.-bab5.pdf Download (120kB)
	Text (Daftar Pustaka) 22081010161.-daftarpustaka.pdf Download (153kB)
	Text (Lampiran) 22081010161.-lampiran.pdf Restricted to Repository staff only Download (805kB)

Abstract

Cyberbullying has become an increasingly prevalent issue due to the rapid growth of social media use. Although Instagram has implemented various protective features, the available mechanisms are not yet fully capable of protecting users from the risks of cyberbullying, particularly within comment sections that enable direct interaction. At the same time, the increasing prevalence of mixed Indonesian and English language usage on social media adds complexity to the cyberbullying detection process. In response to these challenges, this study proposes the development of a browser extension capable of automatically detecting and filtering cyberbullying-related comments in Instagram using the Bidirectional Long Short-Term Memory (BiLSTM) method integrated with Local Interpretable Model-Agnostic Explanations (LIME). This study utilizes a dataset gathered from multiple sources, comprising 25,276 Indonesian and English text samples classified into two classes, normal and bullying. The experimental results show that the best performing model attained an accuracy of 95.97%, a precision of 95.98%, a recall of 95.97%, and an F1-score of 95.96%. These findings demonstrate that the BiLSTM model can effectively distinguish between normal and bullying comments while maintaining high classification accuracy and strong generalization across bilingual datasets. Furthermore, the integration of LIME enables the system to provide explanations for its predictions by highlighting the words or phrases that contribute most significantly to the prediction process. Therefore, the proposed system not only performs automates cyberbullying detection but also improves the transparency and interpretability of its predictions for users.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Sari, Anggraini Puspita	NIDN0716088605	anggraini.puspita.if@upnjatim.ac.id
Thesis advisor	Via, Yisti Vita	NIDN0025048602	yistivia.if@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA75 Electronic computers. Computer science

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

M. Daffa Faiq Husin Syahputra

Date Deposited:

22 Jul 2026 06:29

Last Modified:

22 Jul 2026 07:15

URI:

https://repository.upnjatim.ac.id/id/eprint/57258