Garuda - Garba Rujukan Digital

Pengembangan Sistem Deteksi Objek Botol Real-Time dengan YOLOv8 untuk Aplikasi Vision Triyanto, Dedi; Zidan, Muhammad; Wahyudi, Mochamad; Pujiastuti, Lise; Sumanto, Sumanto
Indonesian Journal Computer Science Vol. 3 No. 1 (2024): April 2024
Publisher : LPPM Universitas Bina Sarana Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31294/ijcs.v3i1.6070

Plastik daur ulang berperan penting dalam menanggulangi masalah limbah lingkungan sekaligus mendukung praktik keberlanjutan. Penelitian ini bertujuan mengembangkan sistem deteksi botol plastik dan kaleng daur ulang secara real-time menggunakan algoritma YOLOv8 yang terkenal akan kecepatan dan akurasinya. Dengan memanfaatkan dataset yang terdiri dari 2.900 gambar dan melatih model melalui Google Colab selama 25 epoch, penelitian ini berhasil menunjukkan performa luar biasa dari YOLOv8, dengan hasil mAP sebesar 99,5%, precision 99,7%, dan recall 99,5%. Model ini terbukti sangat efektif dalam mendeteksi objek daur ulang, memberikan prediksi yang tepat tanpa kesalahan negatif pada confusion matrix. Untuk penelitian lanjutan, disarankan menambah variasi kelas objek seperti botol kaca dan karet serta memperluas dataset guna meningkatkan generalisasi model. Selain itu, pengujian dalam kondisi nyata sangat diperlukan untuk memastikan kinerja optimal dalam lingkungan yang lebih kompleks. Pendekatan serupa dalam penelitian sebelumnya juga telah membuktikan kinerja unggul dalam deteksi real-time, menjadikan metode ini salah satu yang terdepan dalam pengembangan teknologi berbasis YOLO.

Support Vector Machine Untuk Klasifikasi Penyakit Diabetes Mellitus Triyanto, Dedi
Media Teknologi dan Informatika Vol. 1 No. 3 (2024): Juli
Publisher : Universitas Bina Sarana Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31294/mti.v1i3.6267

Diabetes mellitus adalah penyakit metabolik kronis yang semakin umum dan dapat menyebabkan komplikasi serius jika tidak didiagnosis dan diklasifikasikan dengan tepat. Salah satu masalah utama yang dihadapi adalah kurangnya metode diagnosis yang cepat dan akurat untuk mendeteksi diabetes pada tahap awal. Pendekatan manual seringkali memakan waktu dan tidak memberikan hasil yang optimal, sehingga diperlukan metode berbasis teknologi yang lebih efisien. Penelitian ini bertujuan untuk mengembangkan model pembelajaran mesin menggunakan algoritma Support Vector Machine (SVM) yang dapat membantu dalam klasifikasi diabetes mellitus dengan akurasi yang lebih tinggi. Data yang digunakan berasal dari database kasus diabetes yang mencakup berbagai parameter klinis pasien. Sebelum data digunakan untuk pelatihan model, dilakukan pra-pemrosesan yang mencakup analisis komponen utama dan normalisasi, sehingga fitur yang paling relevan dapat dipilih. Model SVM kemudian digunakan untuk melakukan klasifikasi biner, yaitu menentukan apakah seorang pasien memiliki diabetes mellitus atau tidak. Untuk mengevaluasi kinerja model ini, digunakan beberapa metrik, termasuk akurasi, presisi, recall, dan skor f1. Hasil penelitian menunjukkan bahwa model SVM mencapai tingkat akurasi sebesar 88%, yang menandakan bahwa algoritma ini memiliki potensi besar dalam membantu proses diagnosis diabetes dengan cepat dan akurat. Dengan demikian, model ini diharapkan dapat menjadi solusi dalam mengatasi keterbatasan metode diagnosis tradisional, serta membantu tenaga medis dalam memberikan diagnosis yang lebih tepat, sehingga komplikasi yang diakibatkan oleh diabetes dapat dicegah lebih efektif. Diabetes mellitus is a chronic metabolic disease that is becoming increasingly common and can lead to serious complications if not diagnosed and classified correctly. One of the major challenges faced is the lack of fast and accurate diagnostic methods for detecting diabetes at an early stage. Manual approaches are often time-consuming and do not provide optimal results, highlighting the need for more efficient, technology-based methods. This study aims to develop a machine learning model using the Support Vector Machine (SVM) algorithm to assist in the classification of diabetes mellitus with higher accuracy. The data used in this study comes from a diabetes case database containing various clinical parameters of patients. Before the data is used for model training, pre-processing steps are performed, including principal component analysis and normalization, to select the most relevant features. The SVM model is then applied to perform binary classification, determining whether or not a patient has diabetes mellitus. Several metrics, including accuracy, precision, recall, and F1-score, are used to evaluate the model’s performance. The results show that the SVM model achieves an accuracy rate of 88%, indicating that this algorithm has great potential to help in diagnosing diabetes quickly and accurately. Thus, this model is expected to be a solution to overcome the limitations of traditional diagnostic methods and to assist medical professionals in providing more precise diagnoses, thereby preventing complications caused by diabetes more effectively.

Strengthening MSME Marketing through Digital Marketing and Financial Management Training: International Community Service in Cianjur, West Jawa Umiyati, Hesti; Tribuana, Dhimas; Hermawan, Budi; Suherman, Eman; Triyanto, Dedi; Dewi, Puri Swastika Gusti Krisna; Whardani, Kristina Whardani
MOVE: Journal of Community Service and Engagement Vol. 4 No. 5 (2025): May 2025
Publisher : EQUATOR SINAR AKADEMIA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54408/move.v4i5.450

This international community service, organized by the Doctoral Students of the Management Science Doctoral Program, Class of 2023, at UNIKOM, aimed to enhance the marketing and managerial capabilities of Micro, Small, and Medium Enterprises (MSMEs) affiliated with the Bumi Cianjur Cooperative in West Java. Conducted on February 16–17, 2025, the event was attended by 10 MSMEs operating in various sectors, including services, consumer goods, marketing, production, and savings and loans. The activities included an MSME product exhibition, digital marketing training, financial management workshops, and business consultations. Distinguished speakers from both Indonesia and abroad contributed to the program. The results showed improvements in the participants’ understanding of digital marketing strategies and financial planning, contributing to increased business sustainability and competitiveness. This initiative demonstrates the value of academic-community collaboration and the role of higher education in empowering local enterprises through international engagement.

Pemilihan Provider Internet Ponsel Terbaik Dengan Metode Weighted Sum Model Dyas, Pramudya Widyastama; Purnama, Rachmat Adi; Triyanto, Dedi; Kurniawan, Deny
Media Teknologi dan Informatika Vol. 1 No. 4 (2024): Oktober
Publisher : Universitas Bina Sarana Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31294/mti.v1i4.7623

Penelitian ini bertujuan untuk menganalisis pemilihan provider jaringan ponsel terbaik berdasarkan pengalaman pelanggan menggunakan metode Weighted sum model (WSM). Enam provider utama yang diteliti adalah XL Axiata, Telkomsel, Three, Smartfren, Indosat, dan Axis. Data dikumpulkan melalui kuesioner dengan skala Likert 1-5, di mana nilai 1 menunjukkan "Sangat Tidak Baik" dan nilai 5 menunjukkan "Sangat Baik." Analisis dilakukan secara manual dan dengan menggunakan MATLAB untuk memastikan validitas hasil. Hasil penelitian menunjukkan bahwa Telkomsel mendapatkan skor tertinggi sebesar 205,45, mengindikasikan bahwa provider ini memberikan pengalaman pelanggan terbaik dibandingkan dengan provider lainnya. Faktor-faktor yang paling mempengaruhi kepuasan pelanggan meliputi kualitas sinyal, kecepatan internet, cakupan geografis, harga layanan, dan kualitas layanan pelanggan. Konsistensi antara perhitungan manual dan MATLAB menunjukkan keandalan metode WSM yang digunakan. Penelitian ini memberikan wawasan berharga bagi provider untuk meningkatkan kualitas layanan mereka dan membantu konsumen dalam membuat keputusan yang lebih terinformasi. Saran diberikan kepada Telkomsel untuk mempertahankan kualitas layanannya, sementara provider lain disarankan untuk meningkatkan aspek-aspek layanan yang kurang memuaskan. Penelitian ini juga membuka peluang bagi penelitian lanjutan untuk mencakup lebih banyak variabel dan provider. This study aims to analyze the selection of the best mobile internet provider based on customer experience using the Weighted sum model (WSM). The six main providers studied are XL Axiata, Telkomsel, Three, Smartfren, Indosat, and Axis. Data were collected through a questionnaire using a Likert scale of 1-5, where 1 indicates "Very Poor" and 5 indicates "Very Good." The analysis was conducted manually and using MATLAB to ensure result validity. The results showed that Telkomsel obtained the highest score of 205.45, indicating that this provider offers the best customer experience compared to other providers. Factors that most influence customer satisfaction include signal quality, internet speed, geographic coverage, service price, and customer service quality. The consistency between manual calculations and MATLAB shows the reliability of the WSM method used. This research provides valuable insights for providers to improve their service quality and helps consumers make more informed decisions. Recommendations are given to Telkomsel to maintain its service quality, while other providers are advised to improve aspects of their services that are less satisfactory. This research also opens opportunities for further studies to include more variables and providers.

Support Vector Machine Untuk Klasifikasi Penyakit Diabetes Mellitus Triyanto, Dedi
Media Teknologi dan Informatika Vol. 1 No. 3 (2024): Juli
Publisher : Universitas Bina Sarana Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31294/mti.v1i3.6267

Diabetes mellitus adalah penyakit metabolik kronis yang semakin umum dan dapat menyebabkan komplikasi serius jika tidak didiagnosis dan diklasifikasikan dengan tepat. Salah satu masalah utama yang dihadapi adalah kurangnya metode diagnosis yang cepat dan akurat untuk mendeteksi diabetes pada tahap awal. Pendekatan manual seringkali memakan waktu dan tidak memberikan hasil yang optimal, sehingga diperlukan metode berbasis teknologi yang lebih efisien. Penelitian ini bertujuan untuk mengembangkan model pembelajaran mesin menggunakan algoritma Support Vector Machine (SVM) yang dapat membantu dalam klasifikasi diabetes mellitus dengan akurasi yang lebih tinggi. Data yang digunakan berasal dari database kasus diabetes yang mencakup berbagai parameter klinis pasien. Sebelum data digunakan untuk pelatihan model, dilakukan pra-pemrosesan yang mencakup analisis komponen utama dan normalisasi, sehingga fitur yang paling relevan dapat dipilih. Model SVM kemudian digunakan untuk melakukan klasifikasi biner, yaitu menentukan apakah seorang pasien memiliki diabetes mellitus atau tidak. Untuk mengevaluasi kinerja model ini, digunakan beberapa metrik, termasuk akurasi, presisi, recall, dan skor f1. Hasil penelitian menunjukkan bahwa model SVM mencapai tingkat akurasi sebesar 88%, yang menandakan bahwa algoritma ini memiliki potensi besar dalam membantu proses diagnosis diabetes dengan cepat dan akurat. Dengan demikian, model ini diharapkan dapat menjadi solusi dalam mengatasi keterbatasan metode diagnosis tradisional, serta membantu tenaga medis dalam memberikan diagnosis yang lebih tepat, sehingga komplikasi yang diakibatkan oleh diabetes dapat dicegah lebih efektif. Diabetes mellitus is a chronic metabolic disease that is becoming increasingly common and can lead to serious complications if not diagnosed and classified correctly. One of the major challenges faced is the lack of fast and accurate diagnostic methods for detecting diabetes at an early stage. Manual approaches are often time-consuming and do not provide optimal results, highlighting the need for more efficient, technology-based methods. This study aims to develop a machine learning model using the Support Vector Machine (SVM) algorithm to assist in the classification of diabetes mellitus with higher accuracy. The data used in this study comes from a diabetes case database containing various clinical parameters of patients. Before the data is used for model training, pre-processing steps are performed, including principal component analysis and normalization, to select the most relevant features. The SVM model is then applied to perform binary classification, determining whether or not a patient has diabetes mellitus. Several metrics, including accuracy, precision, recall, and F1-score, are used to evaluate the model’s performance. The results show that the SVM model achieves an accuracy rate of 88%, indicating that this algorithm has great potential to help in diagnosing diabetes quickly and accurately. Thus, this model is expected to be a solution to overcome the limitations of traditional diagnostic methods and to assist medical professionals in providing more precise diagnoses, thereby preventing complications caused by diabetes more effectively.

Pengembangan Sistem Deteksi Objek Botol Real-Time dengan YOLOv8 untuk Aplikasi Vision Triyanto, Dedi; Zidan, Muhammad; Wahyudi, Mochamad; Pujiastuti, Lise; Sumanto, Sumanto
Indonesian Journal Computer Science Vol. 3 No. 1 (2024): April 2024
Publisher : LPPM Universitas Bina Sarana Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31294/ijcs.v3i1.6070

Plastik daur ulang berperan penting dalam menanggulangi masalah limbah lingkungan sekaligus mendukung praktik keberlanjutan. Penelitian ini bertujuan mengembangkan sistem deteksi botol plastik dan kaleng daur ulang secara real-time menggunakan algoritma YOLOv8 yang terkenal akan kecepatan dan akurasinya. Dengan memanfaatkan dataset yang terdiri dari 2.900 gambar dan melatih model melalui Google Colab selama 25 epoch, penelitian ini berhasil menunjukkan performa luar biasa dari YOLOv8, dengan hasil mAP sebesar 99,5%, precision 99,7%, dan recall 99,5%. Model ini terbukti sangat efektif dalam mendeteksi objek daur ulang, memberikan prediksi yang tepat tanpa kesalahan negatif pada confusion matrix. Untuk penelitian lanjutan, disarankan menambah variasi kelas objek seperti botol kaca dan karet serta memperluas dataset guna meningkatkan generalisasi model. Selain itu, pengujian dalam kondisi nyata sangat diperlukan untuk memastikan kinerja optimal dalam lingkungan yang lebih kompleks. Pendekatan serupa dalam penelitian sebelumnya juga telah membuktikan kinerja unggul dalam deteksi real-time, menjadikan metode ini salah satu yang terdepan dalam pengembangan teknologi berbasis YOLO.

Strengthening MSME Human Resources in Cianjur Through Knowledge Management Market to Boost Creativity and Sales Syafei, M Yani; Narimawati, Umi; Triyanto, Dedi; Hermawan, Budi
MOVE: Journal of Community Service and Engagement Vol. 4 No. 4 (2025): March 2025
Publisher : EQUATOR SINAR AKADEMIA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54408/move.v4i4.433

Micro, Small, and Medium Enterprises (MSMEs) are one of the most significant contributors to Indonesia's economy. In Cianjur Regency, West Java, MSMEs serve as the backbone of the local economy, offering a wide range of products such as handicrafts, traditional foods, and fashion items. However, MSME actors in this region often face significant challenges in terms of business management, limited knowledge of modern strategies, and restricted access to technology and modern markets. To address these issues, the Knowledge Management Market program was introduced as part of the Community Service Week in Cianjur. This initiative aims to strengthen the human resources (HR) of MSMEs through a knowledge-based approach to enhance creativity and sales.

Implementation of the Geometric Mean Multi-Attribute Utility Theory (G-MAUT) in Determining the Best Honorary Employees Setiawansyah, Setiawansyah; Rahmanto, Yuri; Ulum, Faruk; Triyanto, Dedi
Jurnal Ilmiah Computer Science Vol. 3 No. 2 (2025): Volume 3 Number 2 January 2025
Publisher : PT. SNN MEDIA TECH PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58602/jics.v3i2.50

Determining the best honorary employees is a strategic step to appreciate performance, increase motivation, and encourage productivity in the work environment. This process is carried out by evaluating employees based on certain criteria. The main problem in determining the best honorary employees is the lack of objectivity and transparency in the assessment process, which often leads to dissatisfaction among employees. Judgments that rely solely on subjective perceptions without considering measurable quantitative data can result in unfair decisions. The purpose of applying the Geometric Mean Multi-Attribute Utility Theory (G-MAUT) method in determining the best honorary employees is to provide a more objective, transparent, and accurate evaluation framework in decision-making. This method not only supports a fairer selection process, but also encourages increased motivation and performance among honorary employees. The results of the calculation of the final utility value carried out using the G-MAUT method, the results of the evaluation of eight honorary employees showed their performance ratings comprehensively. Honorary Employee F has the highest utility value of 0.6399, making it the best honorarium employee among all available alternatives. Followed by Honorary Employee A who was ranked second with a utility value of 0.4685, and Honorary Employee D in third place with a value of 0.3947. These results provide a clear picture of the order of employees based on their performance in various criteria that have been assessed.

Privacy-Preserving machine learning in edge computing environments Kurniawan, Deni; Triyanto, Dedi; Wahyudi, Mochamad; Pujiastuti, Lise
Jurnal Teknik Informatika C.I.T Medicom Vol 15 No 3 (2023): July: Intelligent Decision Support System (IDSS)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/cit.Vol15.2023.621.pp118-125

Edge computing has transformed data processing by moving computation closer to the source, enabling real-time analysis and decision-making. Edge devices are decentralized, which creates privacy and confidentiality concerns, especially when applying machine learning algorithms to sensitive data. Privacy-preserving machine learning methods for edge computing are examined in this research. Federated learning, homomorphic encryption, differential privacy, and secure aggregation are examined as data protection methods for network edge machine learning. A thorough study of these methods shows the challenges of balancing privacy, computational economy, and model correctness. Federated learning has promise for collaborative model training without raw data sharing, but communication overhead and convergence speed remain. A fictional healthcare use case shows how federated learning may be used to train collaborative models across many edge devices while protecting patient data. The case study stresses the necessity for sophisticated optimizations to overcome edge device limits and regulatory compliance. Federated learning algorithms, privacy-preserving procedures, and ethics must be improved, according to the research. Future directions include improving heterogeneous edge algorithms, addressing data ownership and consent ethics, and increasing model decision-making openness. This paper presents essential insights on privacy-preserving machine learning in edge computing and advocates for robust techniques for different edge environments. The paper emphasizes the importance of technological advances, ethical frameworks, and regulatory compliance for secure and privacy-aware machine learning in decentralized edge computing

Explainable artificial intelligence (XAI) for trustworthy decision-making Kurniawan, Deni; Triyanto, Dedi; Wahyudi, Mochamad; Pujiastuti, Lise
Jurnal Teknik Informatika C.I.T Medicom Vol 15 No 5 (2023): November : Intelligent Decision Support System (IDSS)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/cit.Vol15.2023.622.pp240-246

This research delves into the optimization of loan approval decisions by integrating the Trustworthy Decision Making (TDM) framework into a mathematical model. The study aims to strike a balance between maximizing loan approvals and ensuring fairness, transparency, and accountability in AI-driven decision-making processes. Leveraging principles of transparency, fairness, and accountability, the mathematical model seeks to optimize loan approvals while adhering to ethical considerations. The formulation emphasizes the importance of interpretable models to maintain transparency in decision explanations, ensuring alignment with trustworthy AI practices. Implementation results demonstrate the efficacy of the model in achieving a balanced approval rate across demographic groups while providing transparent explanations for decisions. This study highlights the significance of ethical considerations and mathematical formulations in fostering responsible AI implementations. However, continual refinement and adaptation of such models remain essential to align with evolving ethical standards and societal expectations. Overall, this research contributes to the discourse on responsible AI by showcasing a methodological approach that integrates ethical principles and mathematical formulations to promote fairness, transparency, and accountability in AI-driven decision-making.

Title

Found 11 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 11 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 11 Documents
Search