Article Per Year (5 Year)
p-Index From 2021 - 2026
1.311
P-Index
This Author published in this journals
All Journal TELKOMNIKA (Telecommunication Computing Electronics and Control) Journal of Information Technology and Computer Science (JOINTECS) JITK (Jurnal Ilmu Pengetahuan dan Komputer) EDUMATIC: Jurnal Pendidikan Informatika Jurnal Tekno Kompak bit-Tech Information System Journal (INFOS)
I Made Artha Agastya
Universitas AMIKOM Yogyakarta
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Education Electrical & Electronics Engineering Engineering Languange, Linguistic, Communication & Media Library & Information Science Mathematics

Published : 7 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 7 Documents
Search

 1 
PENGARUH STEMMER BAHASA INDONESIA TERHADAP PEFORMA ANALISIS SENTIMEN TERJEMAHAN ULASAN FILM I Made Artha Agastya
Jurnal Tekno Kompak Vol 12, No 1 (2018): Februari
Publisher : Universitas Teknokrat Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33365/jtk.v12i1.70

Abstract

Bahasa Indonesia memiliki banyak variasi akhiran, awalan, dan sisipan. Stemming adalah bagian dari prapengolahan dari analisis sentimen yang mendeteksi dan menghilangkan imbuhan tersebut. Pengaruh dari stemming pada analisis sentimen masih belum jelas karena dataset yang digunakan tidak terdistribusi secara bebas. Untuk mendapatkan pengaruh dari stemming terhadap analisis sentimen maka dilakukan percobaan dengan dataset ulasan film yang sudah diterjemahkan ke Bahasa Indonesia. Stemmer Sastrawi sebagai algoritma stemming terbaru digunakan pada penelitian ini. Dataset dibagi menjadi 5 (lima) kategori yang mana 100 data, 250 data, 500 data, 750 data, dan 1000 data. Hasil yang diperoleh menunjukan bahwa stemmer tidak memberikan peningkatan akurasi yang stabil. Bahkan waktu yang diperlukan untuk menyelesaikan analisis sentimen memerlukan waktu meningkat hingga 310 kali lipat. Kenyataan ini sangat buruk karena stemming dapat mengurangi efisiensi dari analisis sentimen.
Implementasi Deep Learning untuk Klasifikasi Motor Imagery pada Sinyal EEG I Made Artha Agastya; Robert Marco2; Nila Feby Puspitasari
JOINTECS (Journal of Information Technology and Computer Science) Vol 8, No 2 (2024)
Publisher : Universitas Widyagama Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31328/jointecs.v8i2.6413

Abstract

Electroencephalography (EEG) adalah teknik yang digunakan untuk merekam aktivitas listrik otak melalui sensor yang ditempatkan pada kulit kepala. Salah satu area penelitian yang menarik dalam analisis EEG adalah motor imagery (MI), yaitu kemampuan untuk membayangkan suatu gerakan tanpa adanya stimulus visual eksternal. Pengolahan sinyal EEG yang kompleks dalam skenario MI memerlukan pendekatan komputasi yang canggih untuk mengenali pola-pola yang terbentuk selama proses pembayangan tersebut. Penelitian ini bertujuan untuk membandingkan kinerja empat arsitektur deep learning populer—yaitu EEGNet, EEGConformer, EEGInception, dan EEGITNet—dalam mengklasifikasikan data EEG pada konteks motor imagery. Hasil pengujian menunjukkan bahwa EEGConformer dan EEGNet adalah model yang paling efektif, dengan akurasi rata-rata masing-masing sebesar 72,41% dan 71,88%, serta performa yang stabil di berbagai subjek. Di sisi lain, EEGInception dan EEGITNet mencatatkan akurasi yang lebih rendah, terutama EEGInception dengan akurasi rata-rata sebesar 55,59%. Temuan ini mengindikasikan bahwa arsitektur sederhana seperti EEGNet tetap kompetitif, meskipun model yang lebih kompleks seperti EEGConformer memberikan sedikit keunggulan dalam performa. Penelitian ini juga menyoroti pentingnya faktor spesifik-subjek dalam meningkatkan performa model, yang dapat diatasi melalui pendekatan adaptif atau personalisasi. Hasil penelitian ini diharapkan dapat memberikan wawasan lebih mendalam terkait model yang paling akurat dalam tugas motor imagery EEG dan berkontribusi pada pengembangan aplikasi berbasis Brain-Computer Interface (BCI).
FOREST FIRE LOCATION AND TIME RECOGNITION IN SOCIAL MEDIA TEXT USING XLM-ROBERTA Hafidz Sanjaya; Kusrini Kusrini; Kumara Ari Yuana; Arief Setyanto; I Made Artha Agastya; Simone Martin Marotta; José Ramón Martínez Salio
JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer) Vol. 10 No. 4 (2025): JITK Issue May 2025
Publisher : LPPM Nusa Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/jitk.v10i4.6194

Abstract

Forest fires have become a serious global threat, significantly impacting ecosystems, communities, and economies. Although remote sensing technology shows potential, limitations such as time delays, limited sensor coverage, and low resolution reduce its effectiveness for real-time forest fire detection. Additionally, social media can serve as a multimodal sensor, presenting multilingual text data with rapid and global coverage. However, it may encounter challenges in obtaining location and time information on forest fires due to limitations in datasets and model generalization. This study aims to develop a multilingual named entity recognition (NER) model to identify location and time entities of forest fires in social media texts such as tweets. Utilizing a transfer learning approach with the XLM-RoBERTa architecture, fine-tuning was performed using the general-purpose Nergrit corpus dataset containing 19 entities, which were relabeled into 3 main entities to detect location, date, and time entities from tweets. This approach significantly improves the model's ability to generalize to disaster domains across multiple languages and noisy social media texts. With a fine-tuning accuracy of 98.58% and a maximum validation accuracy of 96.50%, the model offers a novel capability for disaster management agencies to detect forest fires in a scalable, globally inclusive manner, enhancing disaster response and mitigation efforts.
A comparative study of mango fruit pest and disease recognition Kusrini Kusrini; Suputa Suputa; Arief Setyanto; I Made Artha Agastya; Herlambang Priantoro; Sofyan Pariyasto
TELKOMNIKA (Telecommunication Computing Electronics and Control) Vol 20, No 6: December 2022
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/telkomnika.v20i6.21783

Abstract

Mango is a popular fruit for local consumption and export commodity. Currently, Indonesian mango export at 37.8 M accounted for 0.115% of world consumption. Pest and disease are the common enemies of mango that degrade the quality of mango yield. Specialized treatment in export destinations such as gamma-ray in Australia, or hot water treatment in Korea, demands pest-free and high-quality products. Artificial intelligence helps to improve mango pest and disease control. This paper compares the deep learning model on mango fruit pests and disease recognition. This research compares Visual Geometry Group 16 (VGG16), residual neural network 50 (ResNet50), InceptionResNet-V2, Inception-V3, and DenseNet architectures to identify pests and diseases on mango fruit. We implement transfer learning, adopt all pre-trained weight parameters from all those architectures, and replace the final layer to adjust the output. All the architectures are re-train and validated using our dataset. The tropical mango dataset is collected and labeled by a subject matter expert. The VGG16 model achieves the top validation and testing accuracy at 89% and 90%, respectively. VGG16 is the shallowest model, with 16 layers; therefore, the model was the smallest size. The testing time is superior to the rest of the experiment at 2 seconds for 130 testing images.
Mitigating Class Imbalance in Indonesian Sarcasm Detection: A Cross-Platform Transformer Study A. Salky Maulana; I Made Artha Agastya
Jurnal Pendidikan Informatika (EDUMATIC) Vol 10 No 1 (2026): Edumatic: Jurnal Pendidikan Informatika
Publisher : Universitas Hamzanwadi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29408/edumatic.v10i1.33724

Abstract

Sarcasm detection in Indonesian social media remains challenging due to implicit pragmatic expressions, severe class imbalance, and strong domain variation across platforms. Unlike prior Indonesian sarcasm studies that predominantly focus on in-domain accuracy using conventional balancing methods, this study provides the first systematic cross-platform analysis of generative data balancing under domain shift. We empirically examine whether GPT-4o based generative balancing improves robustness rather than accuracy-centric evaluation in Transformer-based sarcasm detection. Models trained on Twitter data are evaluated across Twitter, Reddit, and TikTok as an unseen domain. The results show that generative balancing yields limited gains in in-domain evaluation but consistently improves cross-domain robustness by increasing sarcasm recall, particularly for Base models. Notably, XLM-R Base achieves an absolute F1-score improvement of +10.8 points on TikTok, while IndoBERT-Large attains the highest in-domain F1-score of 0.7444. These findings indicate that generative augmentation partially mitigates class imbalance by enhancing robustness under domain shift, thereby repositioning sarcasm detection as a robustness-oriented problem and highlighting generative balancing as a complementary strategy rather than a substitute for larger Transformer models in cross-platform NLP settings.
DETEKSI PENYAKIT KULIT DENGAN MENGGUNAKAN MODEL PRETRAINED DAN HYBRID KNOWLEDGE DISTILLATION Theopilus Bayu Sasongko; Arifiyanto Hadinegoro; Eli Pujastuti; I Made Artha Agastya; Nazaruddin Ahmad
Information System Journal Vol. 8 No. 02 (2025): Information System Journal (INFOS)
Publisher : Universitas Amikom Yogyakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24076/infosjournal.2025v8i02.2585

Abstract

Knowledge Distillation (KD) merupakan paradigma efektif untuk mentransfer pengetahuan dari model teacher berkapasitas tinggi ke model student yang ringan melalui kombinasi soft label dan hard label. Meskipun KD Hinton mampu menangkap kesamaan antar kelas, pendekatan ini masih terbatas dalam mentransfer representasi fitur mendalam yang krusial pada tugas pencitraan medis, seperti klasifikasi lesi kulit, di mana fitur halus sering hilang jika hanya mengandalkan keluaran akhir model. Untuk mengatasi keterbatasan tersebut, penelitian ini mengembangkan tiga varian KD, yaitu KD Hinton dengan supervisi hard label, KD dengan penyelarasan fitur, dan Hybrid KD yang mengombinasikan keduanya. Pendekatan ini memungkinkan student meniru distribusi semantik dan representasi fitur internal teacher sekaligus mempertahankan informasi diskriminatif dari ground truth. Eksperimen pada berbagai pasangan teacher–student menunjukkan adanya trade-off antara akurasi dan biaya komputasi. Hasilnya, metode Hybrid KD memberikan peningkatan kinerja tertinggi, mencapai akurasi Top-1 sebesar 82,07% pada MobileNetV2 tanpa menambah kompleksitas model, sehingga efektif untuk aplikasi pencitraan medis real-time berbasis sumber daya terbatas.
Lightweight Model With Hyperparameter Optimization For Classification of Tomato Leaf Diseases Based On Plantvillage Ari Fitriyandhi; Atika Dwi Cahyani; Risca Yunita; Muhammad Ricky Perdana; Taufik Aldri Kristian; Kusrini Kusrini; I Made Artha Agastya
bit-Tech Vol. 8 No. 3 (2026): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i3.3566

Abstract

Tomato cultivation is a vital agricultural commodity in Indonesia, yet leaf diseases continue to pose a serious threat to crop quality and yield. While deep learning–based classifiers have achieved high accuracy in laboratory settings, most existing tomato leaf disease detection models rely on computationally intensive architectures that limit their practical deployment on resource-constrained devices commonly used in agricultural environments. To address this gap, this study proposes a lightweight Convolutional Neural Network (CNN) based on the MobileNetV2 architecture, explicitly combined with systematic hyperparameter optimization, for tomato leaf disease classification. Using 14,529 images from the PlantVillage dataset, the research involves image preprocessing, data augmentation, and structured tuning to improve performance while maintaining computational efficiency. The optimized model achieves an accuracy of 81% using a learning rate of 0.001, 128 units, a dropout rate of 0.3, and an alpha value of 0.35. Although this accuracy is slightly lower than that reported by heavyweight CNN models, it is competitive for lightweight architectures and represents a favorable trade-off between classification performance and computational efficiency. Despite its compact design, the model demonstrates reliable disease recognition and suitability for deployment on devices with limited resources. Furthermore, the trained model was implemented in a desktop-based application as a proof-of-concept system, demonstrating scalability and potential adaptation to mobile or edge-based agricultural decision-support platforms. This study highlights the novelty of integrating lightweight CNN design with systematic hyperparameter optimization and demonstrates that optimized lightweight deep learning models can provide effective, efficient, and deployable solutions for real-world precision agriculture applications.
Co-Authors A. Salky Maulana Ari Fitriyandhi Arief Setyanto Arifiyanto Hadinegoro, Arifiyanto Atika Dwi Cahyani Eli Pujastuti Hafidz Sanjaya, Hafidz Herlambang Priantoro José Ramón Martínez Salio Kumara Ari Yuana Kusrini Muhammad Ricky Perdana Nazaruddin Ahmad Nila Feby Puspitasari Risca Yunita Sasongko, Theopilus Bayu Simone Martin Marotta Sofyan Pariyasto Suputa Suputa Taufik Aldri Kristian