Article Per Year (5 Year)
p-Index From 2021 - 2026
2.715
P-Index
This Author published in this journals
All Journal Jurnal Pendidikan Teknologi dan Kejuruan KARMAPATI (Kumpulan Artikel Mahasiswa Pendidikan Teknik Informatika) ISSN: 2252-9063 Jurnal Sains dan Teknologi Jurnal Penelitian dan Evaluasi pendidikan Jurnal Teknologi Informasi dan Ilmu Komputer Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) SINTECH (Science and Information Technology) Journal JUTIM (Jurnal Teknik Informatika Musirawas) Jurnal Nasional Pendidikan Teknik Informatika (JANAPATI) Locus : Majalah Ilmiah FIA Jnana Karya YASIN: Jurnal Pendidikan dan Sosial Budaya MASALIQ: Jurnal Pendidikan dan Sains Journal of Artificial Intelligence and Digital Business
Agus Seputra I Ketut
Universitas Pendidikan Ganesha
Religion Humanities Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Electrical & Electronics Engineering Engineering Public Health Social Sciences Other

Published : 25 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 25 Documents
Search

 1   2   3 
FINE TUNNING MODEL INDOBERT UNTUK ANALISIS SENTIMEN BERITA PARIWISATA INDONESIA I Nyoman Saputra Wahyu Wijaya; Ketut Agus Seputra; Ni Putu Novita Puspa Dewi
Jurnal Pendidikan Teknologi dan Kejuruan Vol. 22 No. 2 (2025): Edisi Juli 2025
Publisher : Universitas Pendidikan Ganesha

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.23887/jptk-undiksha.v22i2.104056

Abstract

Perkembangan kecerdasan buatan pada ranah NLP dewasa ini sangat pesat. Beberapa teknologi kecerdasan buatan pada text-based teknologi seperti ChatGPT, Gemini, LLaMA, dan lain-lain telah dimanfaatkan dalam ranah riset ataupun industry. Dalam analisis sentimen, yang menjadi komponen utama adalah representasi teks. Teknik representasi teks yang menonjol pada akhir-akhir ini adalah bidirectional encoder representation from transformer(BERT). Sesuai dengan permasalahan yang disebutkan sebelumnya, analisis sentimen ini dapat dilakukan untuk berita pariwisata. Namun untuk meningkatkan akurasi dapat dilakukan fine tunning pada metode BERT. Berdasarkan permasalahan tersebut, dalam penelitian ini akan dilakukan analisis sentiment dengan menggunakan metode IndoBERT. Akan dilakukan fine tuning untuk fokus ranah pariwisata. Berdasarkan hasil pengujian yang telah dilakukan didapatkan tangkat akurasi sebesar 77%. Model dapat melakukan klasifikasi sentiment negative dengan baik, namun masih perlu ditingkatkan pada sentiment positif dan netral.  
Development of A Balinese Tradition and Rite Quiz Game With Text to Speech Feature I Wayan Wahyu Kusuma; I Gede Partha Sindu; Ketut Agus Seputra; Putu Zasya Eka Satya Nugraha; Helmi Maulana Hadiwinata
KARMAPATI (Kumpulan Artikel Mahasiswa Pendidikan Teknik Informatika) Vol. 14 No. 3 (2025)
Publisher : Universitas Pendidikan Ganesha

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.23887/karmapati.v14i3.103037

Abstract

This research aims to develop an educational quiz-based game that presents materials on Balinese traditions and rites, equipped with an automatic question reading feature using a Natural Language Processing (NLP) approach. The method used is Research and Development (R&D) with a one-cycle Agile development approach. The application was developed using Unity, with the integration of Text-To-Speech (TTS) technology through the Tacotron 2 model to automatically generate speech from question text. Testing was conducted through four stages: functional testing using the Unity Test Framework, validation by media experts, validation by content experts, and user response testing using the User Experience Questionnaire – Short Version (UEQ-S). The results showed that the application functioned properly (100% passed), received a highly feasible rating from media experts (90%), and a very high content validity score from content experts (1.00). Meanwhile, user testing resulted in an average score of 2.33, falling into the “Excellent” category. Based on these results, it can be concluded that this educational game application is feasible to use as an interactive learning medium that is both engaging and capable of providing a positive learning experience for users.
Development of the Berry Happy Application for the Digitalization of Bedugul Culinary MSMEs Using the Scrum Approach Ammulia Rizqie Ramadhana Pujiyanto; Putu Nadia Prameswari; Ketut Dian Suryasih; Putu Melianti Eka Maharani; Ketut Agus Seputra; Kadek Teguh Dermawan
RIGGS: Journal of Artificial Intelligence and Digital Business Vol. 5 No. 2 (2026): Mei-Juli
Publisher : Prodi Bisnis Digital Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/riggs.v5i2.9652

Abstract

The High strawberry production in the Bedugul area often results in surplus harvests that are not fully absorbed by the market, creating challenges for local farmers and strawberry-based culinary MSMEs. One of these enterprises is Berry Happy, which processes strawberries into various dessert products. However, the business still relies on manual transaction recording and has limited market reach. This study aims to design and develop a mobile-based e-commerce application to support the digitalization of business processes at Berry Happy. The application was developed using the Flutter framework and the Agile Scrum methodology, which consists of product backlog determination, sprint planning, sprint implementation, sprint review, and sprint retrospective activities. The resulting system provides features for customer registration, product browsing, shopping cart management, checkout processing, transaction history, and menu management for business owners. Functional testing was conducted using the Black Box Testing method to evaluate whether each feature operated according to the specified requirements. The testing results showed that all defined test scenarios were successfully executed and produced outputs consistent with the expected results. The developed application facilitates digital transaction management, supports product promotion through a mobile platform, and provides a more structured approach to recording business transactions. These findings indicate that the application can serve as a practical digital solution to support the operational activities of culinary MSMEs.
Implementation of MQTT Broker and Gemini API in an Internet of Things Based Indoor Air Pollution Monitoring System I Putu Tude Rama Prasatya; Ketut Agus Seputra; Kadek Yota Ernanda Aryanto
RIGGS: Journal of Artificial Intelligence and Digital Business Vol. 5 No. 2 (2026): Mei-Juli
Publisher : Prodi Bisnis Digital Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/riggs.v5i2.10874

Abstract

Indoor air pollution from pollutants such as Carbon Monoxide (CO) and Particulate Matter (PM2.5) poses significant health risks to room occupants, particularly in enclosed spaces with poor ventilation where pollutants can accumulate to hazardous concentrations. To address this challenge, this study designed and implemented an IoT-based indoor air quality monitoring system integrating an MQTT broker for real-time data transmission and the Gemini API for intelligent data interpretation. The system adopts a three-layer architecture spanning hardware, backend, and application layers. The hardware layer utilizes an ESP32 as a wireless gateway and an Arduino Nano for sensor acquisition, employing the Sharp GP2Y1010AU0F sensor for particulate matter and the MQ-7 sensor for carbon monoxide. The backend, built with Laravel, manages data through a dual-database approach, where MySQL handles structured user data and InfluxDB stores continuous sensor readings. A Flutter mobile application, built with the MVVM pattern, serves as the user interface, delivering real-time air quality information. The Gemini API further enhances the system by automatically generating air quality classifications and actionable health recommendations. Calibration testing demonstrated an average error of 6.45% for the MQ-7 sensor and a notably low 0.44% error for the Sharp GP2Y1010AU0F sensor, indicating high measurement accuracy. A 24-hour continuous stress test revealed a system uptime of 83.3%, confirming reasonable operational reliability. Finally, a usability evaluation using the SUS method involving 30 respondents yielded an average score of 70.67, placing the system in Grade B with a "Good" interpretation, confirming the system is functional and easy to use.
Pengembangan Aplikasi Mobile Klasifikasi Object Detection Sampah Anorganik Menggunakan Arsitektur SSD MobilenetV2 Ngakan Gde Satria Abirama; Ketut Agus Seputra; I Nyoman Saputra Wahyu Wijaya
RIGGS: Journal of Artificial Intelligence and Digital Business Vol. 5 No. 2 (2026): Mei-Juli
Publisher : Prodi Bisnis Digital Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/riggs.v5i2.11992

Abstract

Pengelolaan sampah anorganik yang masih dilakukan secara manual menyebabkan proses pemilahan menjadi kurang efisien dan berpotensi meningkatkan kesalahan identifikasi jenis sampah serta Kondisi tersebut dapat menghambat upaya pengelolaan dan daur ulang sampah yang efektif. Penelitian ini bertujuan mengembangkan aplikasi mobile dengan menanamkan model SSD-MobileNetV2 yang mampu mendeteksi sampah plastik dan kaca secara real-time untuk membantu proses identifikasi dan pemilahan sampah plastik dan kaca secara lebih cepat dan akurat.. Penelitian menggunakan metode Research and Development (R&D) yang meliputi tahapan pengumpulan dan pengolahan data, pengembangan model, pengembangan aplikasi, serta pengujian aplikasi. Model dilatih menggunakan pendekatan transfer learning dengan partial fine-tuning pada 10 layer terakhir backbone MobileNetV2 dan diimplementasikan ke dalam aplikasi Android menggunakan TensorFlow Lite sehingga dapat dijalankan secara optimal pada perangkat Android tanpa memerlukan koneksi internet. Berdasarkan hasil evaluasi, model memperoleh akurasi sebesar 89,90%, presisi 90,00%, recall 90,00%, F1-score 90,00%, dan mean Average Precision (mAP) sebesar 95,88%. Hasil pengujian Black Box Testing menunjukkan seluruh fitur aplikasi berjalan sesuai dengan kebutuhan sistem, sedangkan pengujian System Usability Scale (SUS) menunjukkan aplikasi mudah dipahami dan digunakan oleh pengguna. Hasil penelitian menunjukkan bahwa aplikasi yang dikembangkan mampu mendeteksi sampah plastik dan kaca secara cepat dan akurat serta berpotensi menjadi solusi pendukung pemilahan sampah anorganik berbasis perangkat mobile untuk meningkatkan efektivitas pengelolaan sampah.
Co-Authors A. A. Gede Yudhi Paramartha Agus Aan Jiwa Permana Ainan Fajar Fatcha Ammulia Rizqie Ramadhana Pujiyanto Anak Agung Istri Ngurah Marhaeni Dermawan, Kadek Teguh Gede Aditra Pradnyana Gede Arna Jude Saskara Gede Sandiasa Helmi Maulana Hadiwinata I Ketut Resika Arthana I Nyoman Saputra Wahyu Wijaya I Putu Dion Arditya I Putu Tude Rama Prasatya I Wayan Wahyu Kusuma Joe Aqilla Vandyta Kadek Teguh Dermawan Kadek Yota Ernanda Aryanto Ketut Dian Suryasih Komang Diva Andi Wirawan Luh Joni Erawati Dewi Nanda, I Kadek Dwi Angga Sathya Ngakan Gde Satria Abirama Ni Ketut Suarni Ni Ketut Sudianing Ni Putu Novita Puspa Dewi Ni Wayan Marti Ony Andewi, Putu Putu Melianti Eka Maharani Putu Nadia Prameswari Putu Zasya Eka Satya Nugraha Sindu, I Gede Partha Wijaya, I Nyoman Saputra Wahyu