cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Dina Angela
Contact Email
dina_angela@ithb.ac.id
Phone
-
Journal Mail Official
jurnaltelematika@ithb.ac.id
Editorial Address
Jl. Dipati Ukur no. 80-84, Kel. Coblong, Kec. Lebak Gede, Bandung, 40132
Location
Kota bandung,
Jawa barat
INDONESIA
Jurnal Telematika
Published by Institut Teknologi Harapan Bangsa
ISSN : 18582516     EISSN : 25793772     DOI : https://doi.org/10.61769/telematika
Core Subject : Science, Engineering,
Computer Science & IT Electrical & Electronics Engineering
Jurnal Telematika is a scientific periodical written in Indonesian language published by Institut Teknologi Harapan Bangsa twice per year. Jurnal Telematika publishes scientific papers from researchers, academics, activist, and practicioners, which are results from scientific study and research in the field of telematics and information technology.
Arjuna Subject : -
Articles 248 Documents
 21   22   23   24   25 
Klasifikasi Terawasi Anomali Suara Kipas Industri Menggunakan Jaringan Saraf Tiruan dan Fitur Akustik Rekayasa Thenata, Angelina Pramana; ., Ranny; Hakim, Bhustomy; Kaunang, Fergie Joanda
Jurnal Telematika Vol. 20 No. 1 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i1.772

Abstract

Penelitian ini menggunakan pendekatan supervised learning berbasis jaringan saraf untuk deteksi anomali pada sistem kipas industry. Dengan subset data FAN dari MIMII (malfunctioning industrial machine investigation and inspection) dataset dengan 530 rekaman berlabel (383 normal dan 147 abnormal), penelitian ini mengekstraksi fitur akustik yang meliputi mel-frequency cepstral coefficients (MFCC), spectral descriptor (centroid, roll off), serta temporal measures (zero-crossing rate, autocorrelation). Uji statistik univariat menunjukkan sejumlah koefisien MFCC dan fitur domain waktu berbeda signifikan antar kelas (p < 0,05). Model jaringan saraf feed-forward dengan dua lapisan tersembunyi berukuran 64 unit (aktivasi ReLU) dan regularisasi dropout dilatih menggunakan stratified cross validation dengan 5-fold sehingga menghasilkan nilai F1 rata-rata sebesar 89,9%. Penggunaan beberapa nilai ambang (τ ∈ {0,3–0,7}) menegaskan kekokohan model yang terlihat pada hasil data uji dengan nilai ambang terpilih adalah τ = 0,5 yang mencapai precision sebesar 100%, recall = 93,10%, F1 = 96,43%, dan akurasi = 98,11% (hasil identik diperoleh pada τ = 0,6–0,7; sementara τ = 0,3 memberikan recall lebih tinggi). Model juga menghasilkan nilai AUC-ROC sebesar 0,9978 yang mendekati ideal dan menunjukkan daya diskriminasi lintas-ambang yang sangat baik. Temuan ini memperlihatkan bahwa penggabungan fitur akustik yang dapat diinterpretasikan dengan pengklasifikasi saraf yang ringkas memungkinkan deteksi anomali non-invasif yang akurat untuk penerapan Industri 4.0 dengan kebutuhan perangkat keras minimal.
Analisis Watermarking Citra Digital Berbasis Contourlet Transform dan DCT dalam Ruang Warna RGB, YCbCr, dan YIQ Daniel Setiadikarunia; Agus Setiawan
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.741

Abstract

This article discusses watermarking techniques for contourlet transform (CT)-based digital images, combined with the discrete cosine transform (DCT). Watermarks are inserted into one of the colour components of the host image's colour space, namely RGB, YCbCr, or YIQ. Level 2 contourlet transformation is applied to the colour component used for watermark insertion. The upper-right subband of the contourlet transform result is selected, divided into 4×4 blocks, and then DCT and zigzag scanning are applied to each block. The watermark used is a black-and-white (binary) image. To improve security and reduce spatial correlation, the watermark is scrambled using Arnold scrambling. The watermark is embedded in the host image by inserting a PN sequence corresponding to each watermark bit into the midband DCT coefficients of the 4×4 block. Test results show that the inserted image has a high level of imperceptibility (PSNR> 30 dB). The watermark is robust against JPEG compression with a minimum quality factor of 8, scaling of 75% and 125%, rotation of 180°, cropping of 20% with different cropping positions, and the addition of up to 5% Gaussian noise, but is not robust against 90° rotation and median filtering. Watermarks embedded in the YCbCr colour space in the Cb or Cr components, or in the YIQ colour space in the I or Q components, achieve optimal imperceptibility and robustness.
Analisis Pengaruh Karakteristik Masukan Teks terhadap Kinerja MiniLMv2-L6-H384 dan BERT-Base-Uncased pada Quora Question Pairs Ken Ratri Wardani; Inge Martina; Jimmy Fong Xin Wern
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.775

Abstract

Knowledge distillation merupakan teknik untuk menyederhanakan model bahasa besar menjadi model yang lebih ringkas dengan tetap mempertahankan akurasi. Bidirectional encoder representations from transformers (BERT) menawarkan kinerja kuat namun memerlukan sumber daya komputasi besar, sedangkan mini language model (MiniLM) memiliki ukuran model lima kali lebih kecil. Penelitian ini bertujuan membandingkan kinerja kedua model tersebut pada dataset quora question pairs dengan fokus pada pengaruh sequence length dan kelangkaan token (token rarity) terhadap akurasi klasifikasi. Kedua model dilatih menggunakan parameter pelatihan yang identik. Hasil pengujian menunjukkan BERT mencapai akurasi 91,22% dan F1-score 88,17%, sedikit lebih unggul dari MiniLM yang mencapai akurasi 90,12% dan F1-score 86,73%. Namun, MiniLM memberikan kecepatan inferensi 5,3 kali lebih cepat. Temuan ini memberikan panduan empiris untuk optimalisasi model untuk lingkungan dengan keterbatasan sumber daya komputasi atau kebutuhan respons real-time, di mana efisiensi MiniLM dapat diterima dengan sedikit penurunan akurasi. Penelitian mendatang disarankan untuk mengeksplorasi sistem hibrida yang mengalihkan tugas kompleks ke model besar dan tugas umum ke model yang lebih kecil.
Kerangka Decision Support System Berbasis Aturan untuk Evaluasi Izin Mendirikan Bangunan di Kawasan Rawan Longsor dan Banjir Kasus Kabupaten Bogor Dyah Sulistyowati Rahayu; Nenden Nuraini; Zidan Wali Arubusman; Chaerul Cahyadi
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.786

Abstract

Kabupaten Bogor merupakan wilayah dengan tingkat kerentanan tinggi terhadap bencana longsor dan banjir akibat kondisi geomorfologi, curah hujan yang intens, serta tekanan pembangunan yang meningkat. Permintaan izin mendirikan bangunan (IMB/PBG) di daerah ini menimbulkan dilema kebijakan karena pembangunan hunian di zona rawan berpotensi meningkatkan risiko kerugian. Penelitian ini bertujuan menyusun kerangka decision support system (DSS) yang dapat membantu pemerintah dalam mengevaluasi kelayakan izin mendirikan bangunan hunian secara objektif, transparan, dan berbasis risiko. Metode penelitian menggunakan pendekatan studi literatur dengan menelaah kriteria kelayakan yang relevan, meliputi faktor geoteknik, hidrologi serta aspek regulasi. Hasil sintesis menghasilkan kerangka DSS berbasis aturan (rule-based decision support system) dengan tiga keluaran keputusan: Go (layak), Mitigasi Khusus (izin bersyarat dengan syarat teknis seperti rumah panggung, fondasi tiang, atau drainase lereng), dan No-Go (tidak layak). Kontribusi penelitian ini adalah memberikan model konseptual DSS yang dapat diintegrasikan dengan Sistem Perizinan Bangunan (SIMBG) sehingga mendukung konsistensi kebijakan, memperkuat mitigasi bencana, dan meningkatkan akuntabilitas tata ruang di Kabupaten Bogor.
Pemetaan dan Analisis Sebaran Tanaman Tembakau Berbasis Sistem Informasi Geografis Defilia Fatikasari; Ary Iswahyudi
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.794

Abstract

Permasalahan utama penelitian ini adalah kurangnya pemetaan dan analisis sebaran tanaman tembakau di Kecamatan Proppo, Kabupaten Pamekasan, yang akurat untuk mendukung perencanaan pertanian dan pengelolaan lahan. Penelitian ini bertujuan untuk memetakan dan menganalisis pola persebaran tanaman tembakau berdasarkan faktor lingkungan dan penggunaan lahan menggunakan pendekatan Sistem Informasi Geografis (SIG). Data atau fitur-fitur yang digunakan meliputi citra satelit dari Google Earth Engine, data spasial administratif, dan atribut lahan. Analisis dilakukan dengan teknik overlay intersect untuk menghitung luas lahan berdasarkan batas administratif serta perhitungan indeks Moran untuk mendeteksi pola spasial distribusi tanaman tembakau. Pengolahan data dilakukan dengan menggunakan perangkat lunak QGIS 3.40. Hasil penelitian menunjukkan bahwa total area tembakau mencapai 3.335,344 hektare dengan pola persebaran yang cenderung acak, terutama di wilayah utara dan timur, meskipun terdapat indikasi awal pengelompokan (clustered) di beberapa wilayah barat dan tengah. Nilai Moran’s I sebesar 0,436 menunjukkan kecenderungan pengelompokan, namun belum signifikan secara statistik. Penelitian ini memberikan kontribusi berupa visualisasi sebaran tanaman tembakau berbasis SIG yang dapat mendukung perencanaan pertanian dan pengelolaan lahan berbasis data spasial.
Sentimen Publik Terhadap Kebijakan Pemindahan Ibu Kota Indonesia di X Menggunakan Model BiLSTM-CNN Wanda Nugraha; Mochamad Tito Julianto; Mohamad Khoirun Najib; Elis Khatizah
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.796

Abstract

The development of Indonesia's new capital city, Ibu Kota Nusantara (IKN), is an innovative government policy that has sparked diverse public responses. This study aims to explore sentiment trends on the social media platform X to understand public perceptions of the policy. Additionally, a sentiment classification model combining Bidirectional Long Short-Term Memory (BiLSTM) and Convolutional Neural Network (CNN) was developed and optimized through hyperparameter tuning. Exploratory analysis showed that positive sentiment dominated at 46%, followed by negative at 30% and neutral at 24%. The classification model achieved a test accuracy of 78% and an average accuracy of 81% across 10-fold cross-validation, with a standard deviation of 0.006. The achieved accuracy, together with the low cross-validation standard deviation, indicates that the BiLSTM-CNN model demonstrates stable and reliable performance.
Studi Perbandingan Model Lightweight YOLOv12 untuk Deteksi Objek Bawah Air Secara Real-Time Hebron Prasetya; Revin R. Balo; Tasya Tumbal; Alwin M. Sambul; Muhamad Dwisnanto Putro
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.799

Abstract

Deep learning methods in computer vision play a crucial role in object localization using camera-based sensors, with Convolutional Neural Networks serving as the dominant approach for object detection. However, many existing models incur high computational costs due to deep architectures and complex operations, limiting their use for real-time deployment on low-cost, resource-constrained devices. The YOLOv12 architecture offers lightweight variants to improve computational efficiency. This study evaluates the trade-off between efficiency and detection performance by comparing model variants using the number of parameters, floating-point operations, and inference speed, while detection accuracy is measured using mean average precision. The results assess the suitability of lightweight models for real-time deployment in resource-constrained environments such as underwater monitoring and conservation. Experimental results on the Real-World Underwater Object Detection dataset demonstrate that YOLOv12-nano achieves 5.7% lower accuracy compared to YOLOv12-medium but requires only 2.57 million parameters and 6.5 GFLOPs, significantly less than YOLOv12-medium with 20.1 million parameters and 67.8 GFLOPs. Moreover, YOLOv12-small requires 9.26 million parameters and 21.5 GFLOPs, positioning it between nano and medium in terms of complexity while still maintaining competitive accuracy. In the inference process, YOLOv12-nano achieves 16.48 FPS on a 12th Gen Intel(R) Core (TM) i5-12450HX CPU. In comparison, YOLOv12-small runs at 6.28 FPS, while YOLOv12-medium runs at 2.36 FPS. These results indicate that YOLOv12-nano is the most suitable variant for real-time deployment on CPU-based platforms.
Conversational Information System Dashboard for Zero New Stunting Campaign in Kecamatan Cikarang Timur Tjong Wan Sen; Yosi Yonata
Jurnal Telematika Vol. 20 No. 2 (2025)
Publisher : Yayasan Petra Harapan Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.61769/telematika.v20i2.734

Abstract

The critical issue of stunting eradication necessitates innovative solutions for effective monitoring, intervention, and prevention. In this paper, we introduce an innovative integrated information system dashboard that leverages the power of data analytics and real-time data processing to provide comprehensive insights into stunting eradication efforts. This assists healthcare providers and policymakers in accessing actionable information crucial for making informed decisions. Our dashboard utilizes digitalization technology to analyze and visualize stunting data in almost real-time, operating nonstop 24/7. This platform enables users to engage in interactive and data-driven decision-making processes. The personalized and user-friendly interface allows for a tailored experience that meets the specific needs and preferences of different stakeholders. We used the Rapid Application Development method with an object-oriented approach to develop this dashboard. We demonstrate the proof of concept of this dashboard through a series of experiments and evaluations. Based on the results of the trial, the designed platform can be very useful for Zero New Stunting campaigns in Kecamatan Cikarang Timur.

Page 25 of 25 | Total Record : 248

 21   22   23   24   25 

Filter by Year

2009 2025


Reset
Filter By Issues
All Issue Vol. 20 No. 2 (2025) Vol. 20 No. 1 (2025) Vol. 19 No. 2 (2024) Vol. 19 No. 1 (2024) Vol. 18 No. 2 (2023) Vol. 18 No. 1 (2023) Vol. 17 No. 2 (2022) Vol. 17 No. 1 (2022) Vol. 16 No. 2 (2021) Vol. 16 No. 1 (2021) Vol. 15 No. 2 (2020) Vol. 15 No. 1 (2020) Vol. 14 No. 2 (2019) Vol. 14 No. 1 (2019) 2018: Industrial Engineering Seminar and Call for Paper (IESC) 2018 Vol. 13 No. 2 (2018) Vol. 13 No. 1 (2018) Vol. 12 No. 2 (2017) Vol. 12 No. 1 (2017) Vol. 11 No. 2 (2016) Vol. 11 No. 1 (2016) Vol. 10 No. 2 (2015) Vol. 10 No. 1 (2015) Vol. 9 No. 2 (2014) Vol. 9 No. 1 (2014) Vol. 8 No. 2 (2013) Vol. 8 No. 1 (2013) Vol. 7 No. 2 (2011) Vol. 7 No. 1 (2011) Vol. 6 No. 1 (2010) Vol. 5 No. 1 (2009) More Issue