Claim Missing Document
Check
Articles

Found 2 Documents
Search

XSentiment-HS: Hierarchical CNN-BiGRU-SVM with Explainable for Indonesian Multi-Level Hate Speech Detection Gina Khayatun Nufus; Rizki Dewantara; Ardi Susanto; Sokid; Lia Farhatuaini; Jaka Septiadi; Mohammad Raihan Akbar
Journal of Applied Informatics Science Volume 2 Issue 1 (2026)
Publisher : GWS Tech Solution

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65897/jais.v2.i1.81

Abstract

Deteksi ujaran kebencian pada media sosial menuntut interpretasi teks yang kompleks karena sifatnya yang spontan dan ambigu, terutama dalam bahasa Indonesia yang kaya akan slang. Tantangan utama saat ini adalah keterbatasan penelitian sebelumnya yang mayoritas hanya melakukan klasifikasi biner tanpa mendeteksi tingkat keparahan konten. Penelitian ini mengusulkan XSentiment-HS, sebuah model deep learning hierarkis dua tahap untuk deteksi multi-tingkat hate speech. Arsitektur model menggabungkan Convolutional Neural Networks (CNN) untuk ekstraksi fitur lokal dan Bidirectional Gated Recurrent Unit (BiGRU) untuk menangkap ketergantungan kontekstual jangka panjang. Model ini juga diperkuat dengan mekanisme Multi-Head Attention dan Support Vector Machine (SVM) sebagai classifier final. Melalui integrasi ini, XSentiment-HS diharapkan mampu mengatasi tantangan ekstraksi fitur dan polisemi secara lebih efektif dibandingkan metode konvensional.
Designing an Interactive Learning Game BUDIAN Based on Culture and Arts with a Design Thinking Approach Muhammad Isfa' Lana; Jaka Septiadi; Rizki Hikmawan; Rani Yuniar; Nabiilah Lintang Dadari
Journal of Educational Sciences Vol. 10 No. 3 (2026): Journal of Educational Sciences
Publisher : FKIP - Universitas Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31258/jes.10.3.p.530-549

Abstract

The purpose of this research is to develop an interactive learning application design, BUDIAN, to expand access to arts and culture learning and address the limitations of arts facilities in schools. The research used the Research and Development (R&D) method with a Design Thinking approach through the stages of empathize, define, ideate, prototype, and test. The development results showed that the initial trial on 13 respondents resulted in an average feature access time of 234.1 seconds and several inputs related to navigation, which were then improved in the revision stage. Further testing using the System Usability Scale (SUS) on 25 respondents showed an average score of 84.3 (Excellent category), indicating the application has a high level of usability. This research has an impact on the availability of interactive, contextual, and user-centered cultural learning application designs, and fills the gap in previous research by comprehensively documenting the integration of Design Thinking in R&D. The BUDIAN application has the potential to increase cultural literacy and support the preservation of Indonesian culture through innovative digital media.