Article Per Year (5 Year)
p-Index From 2021 - 2026
2.195
P-Index
This Author published in this journals
All Journal JOURNAL OF INFORMATION SYSTEM MANAGEMENT (JOISM) G-Tech : Jurnal Teknologi Terapan Journal of Information Technology (JIfoTech) Jurnal Pengabdian West Science Jurnal Komputer Antartika Science Information System and Technology SISFOTENIKA Jurnal Pendidikan dan Pengajaran Guru Sekolah Dasar (JPPGuseda)
Aditya Pratama
Universitas Nahdlatul Ulama Kalimantan Barat
Humanities Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Education Energy Engineering Public Health Social Sciences

Published : 13 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 13 Documents
Search

 1   2 
Forest Fire Detection Based on Digital Imagery Using Convolutional Neural Network (CNN) Model Candra Gudiato; Aditya Pratama; Christian Cahyaningtyas
G-Tech: Jurnal Teknologi Terapan Vol 10 No 2 (2026): G-Tech, Vol. 10 No. 2 April 2026
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/g-tech.v10i2.9422

Abstract

This study explores the implementation of a Convolutional Neural Network (CNN) for automated forest fire identification using digital image processing. Utilizing the USTC 'Forest Fire' dataset, the research framework involved systematic data preprocessing, including a 70:30 training-validation split and the application of image augmentation techniques to enhance model robustness. The proposed architecture features a sequential design with dual convolution and pooling layers, integrated with ReLU and Sigmoid activations. Although initial training over seven epochs yielded a deceptive validation accuracy of 99%, granular performance analysis exposed critical limitations. Evaluation via a Confusion Matrix revealed that while the model excelled at identifying 'non-fire' scenarios, it struggled significantly with actual fire detection, failing to recognize 301 out of 331 fire instances. These results highlight a severe class imbalance issue, suggesting that standard accuracy metrics are insufficient for this application and emphasizing the need for more balanced sampling or advanced architectural adjustments in future fire detection systems.
Klasifikasi Citra Kebakaran Hutan Menggunakan Arsitektur ResNet-50 Berbasis Transfer Learning: Forest Fire Image Classification Using ResNet-50 Architecture Based on Transfer Learning Aditya Pratama; Candra Gudiato; Denny Primanda; Weli Zuandi; Wahyu Prayitno
SISFOTENIKA Vol. 16 No. 2 (2026): SISFOTENIKA
Publisher : STMIK PONTIANAK

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30700/sisfotenika.v16i2.707

Abstract

Kebakaran hutan dan lahan (karhutla) merupakan bencana lingkungan serius yang memerlukan sistem deteksi dini secara akurat dan efisien guna meminimalisir dampak kerusakan. Penelitian ini bertujuan untuk menguji keandalan dan efisiensi arsitektur ResNet-50 berbasis Transfer Learning dalam mengklasifikasikan citra digital kebakaran hutan ke dalam dua kelas biner (wildfire dan nowildfire). Dataset sekunder yang diunduh dari repositori terbuka Kaggle terlebih dahulu melalui tahap pra-pemrosesan, mencakup pemeriksaan integritas data (data integrity check) untuk menangani file gambar yang rusak/korup, pengubahan ukuran citra (resizing) menjadi 128 × 128 piksel, normalisasi nilai piksel, serta teknik augmentasi data acak. Dataset dibagi secara terpisah dengan proporsi rasio 70% data latih (30.250 citra), 15% data validasi (6.300 citra), dan 15% data uji (6.300 citra). Arsitektur ResNet-50 dikombinasikan dengan classification head buatan (Global Average Pooling 2D, Dense Layer 256 unit dengan ReLU, Dropout 0.4, dan Output Sigmoid) serta dilatih selama 5 epoch menggunakan optimizer Adam (default learning rate 1 × 10⁻3) dan fungsi kerugian Binary Cross-Entropy Loss. Evaluasi terhadap 6.300 citra uji independen menunjukkan bahwa model mampu mencapai Akurasi Keseluruhan sebesar 88%. Model terbukti memiliki sensitivitas tinggi terhadap kelas kebakaran (wildfire) dengan Recall sebesar 0.90 dan F1-Score sebesar 0.90 (berhasil mengenali 3.141 dari 3.480 citra kebakaran secara tepat). Pengujian Receiver Operating Characteristic (ROC) Curve menghasilkan nilai Area Under Curve (AUC) sebesar 0.92, yang menegaskan kemampuan pemisahan kelas (class separation ability) yang sangat unggul. Hasil ini membuktikan bahwa pendekatan Transfer Learning berbasis ResNet-50 sangat efektif dan efisien untuk diterapkan sebagai mesin pemrosesan visual utama pada sistem peringatan dini kebakaran hutan
Virtual Mathematics Kits: Interactive Gamification for Numeracy Learning in Elementary Education Suriyana Suriyana; Yunika Afryaningsih; Aditya Pratama
Jurnal Pendidikan dan Pengajaran Guru Sekolah Dasar (JPPGuseda) Vol. 9 No. 2 (2026)
Publisher : Universitas Pakuan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55215/jppguseda.v9i2.50

Abstract

This study aimed to develop an interactive Virtual Mathematics Kit (VMK) integrated with a gamification approach and to examine its feasibility, practicality, and effectiveness in improving elementary school students’ numeracy skills. The study was motivated by persistent low numeracy achievement and disparities in students’ mathematical competencies, particularly in rural elementary schools. A Research and Development (R&D) approach was employed using the ADDIE model, comprising the Analysis, Design, Development, Implementation, and Evaluation stages. The participants were 26 fourth-grade elementary school students. The developed interactive VMK was validated by four expert validators in terms of content and media quality. Data were collected through observations, questionnaires, and pretest–posttest assessments and analyzed using descriptive statistics and normalized gain (N-gain) analysis. The validation results indicated that the VMK achieved a feasible category based on evaluations by both material and media experts. Practicality testing showed mean student response scores of 4.2 in individual trials and 3.6 in small-group trials, both categorized as practical. The effectiveness evaluation produced an average N-gain score of 0.58, indicating a moderate improvement in students’ numeracy skills. These findings demonstrate that the gamification-integrated VMK is feasible, practical, and effective for supporting mathematics learning and reducing numeracy gaps among elementary school students. The study contributes to the development of technology-enhanced, gamified learning media to strengthen numeracy education, particularly in rural elementary school contexts.
Co-Authors Adha Maulana Adha Maulana Awanis Hidayati Cahyaningtyas, Christian Denny Primanda Denny Primanda Gudiato, Candra Hendri Hendri Nurman Hariyanto Nurman Hariyanto Rendy Amy Saputra Suriyana Suriyana Venesa Asri Wahyu Prayitno Wahyu Prayitno Weli Zuandi Yunika Afryaningsih