Hartono, Yoga (2025) Pengembangan Sistem Deteksi Ujaran Kebencian Dan Pelecehan Seksual Berbasis Teks Dengan Otomatisasi Pada Platform Media Sosial “X” Menggunakan Natural Language Processing. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5027211023-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5027211023-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (11MB) | Request a copy |
Abstract
Tingginya penyebaran ujaran kebencian dan pelecehan seksual di platform media sosial, khususnya pada platform “X”, menjadi isu yang semakin mengkhawatirkan karena berdampak negatif terhadap kesehatan psikologis pengguna. Meskipun telah dilakukan berbagai upaya penanganan, sistem moderasi yang ada saat ini masih belum mampu mendeteksi dan mengelola konten berbahaya secara efektif. Penelitian ini mengusulkan pengembangan sistem deteksi ujaran kebencian dan pelecehan seksual berbasis teks menggunakan pendekatan Natural Language Processing (NLP). Model utama yang digunakan adalah IndoBERT, sebuah arsitektur berbasis transformer yang telah di-fine-tuning untuk tugas klasifikasi biner pada teks berbahasa Indonesia, khususnya pada cuitan. Hasil evaluasi komparatif dengan model pembanding BiLSTM menunjukkan bahwa IndoBERT memiliki performa yang lebih unggul, dengan perolehan nilai F1-score tertinggi sebesar 0,9470 pada dataset pelecehan seksual dan 0,853 pada dataset ujaran kebencian. Berdasarkan hasil tersebut, IndoBERT dipilih sebagai model final yang diimplementasikan dalam sistem. Untuk mendukung kendali pengguna, sistem juga dilengkapi fitur moderasi otomatis seperti auto-mute guna memfasilitasi pengelolaan interaksi negatif secara proaktif. Hasil dari penelitian ini diharapkan dapat berkontribusi pada pengembangan alat moderasi cerdas serta terciptanya lingkungan digital yang lebih aman dan inklusif bagi pengguna media sosial.
=======================================================================================================================================
The widespread prevalence of hate speech and sexual harassment on social media platforms, particularly on “X”, has become an increasingly concerning issue due to its negative impact on users' psychological well-being. Despite various mitigation efforts have been implemented, existing moderation systems still fall short in effectively detecting and managing harmful content. This study proposes the development of a text-based detection system for hate speech and sexual harassment using a Natural Language Processing (NLP) approach. The primary model used is IndoBERT, a transformer-based architecture fine-tuned for binary classification tasks on Indonesian-language texts, specifically tweets. Comparative evaluation with a baseline BiLSTM model shows that IndoBERT outperforms with the highest F1-score of 0.9470 on the sexual harassment dataset and 0.853 on the hate speech dataset. Based on these results, IndoBERT was selected as the final model implemented in the system. To support user control, the system is also equipped with automated moderation features such as auto-mute to facilitate proactive management of negative interactions. The outcomes of this research are expected to contribute to the development of intelligent moderation tools and the creation of a safer and more inclusive digital environment for social media users.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | NLP, Ujaran Kebencian, Pelecehan Seksual, IndoBERT, AI, NLP, Hate Speech, Sexual Harassment, IndoBERT, AI. |
| Subjects: | Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.76.A63 Application program interfaces Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9.I52 Information visualization T Technology > T Technology (General) > T57.5 Data Processing |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis |
| Depositing User: | Yoga Hartono |
| Date Deposited: | 27 Jul 2025 01:56 |
| Last Modified: | 27 Jul 2025 01:56 |
| URI: | http://repository.its.ac.id/id/eprint/121724 |
Actions (login required)
 |
View Item |