E-Journal Politeknik Negeri Cilacap
Not a member yet
919 research outputs found
Sort by
Perbandingan Kinerja Model Deep Learning Convolutional Neural Network (CNN) dan Multilayer Perceptron (MLP) untuk Klasifikasi Penyakit Diabetes Melitus
No full textDiabetes mellitus is a chronic disease with a continuously increasing number of sufferers. Early detection remains difficult because conventional methods often only recognize the disease at an advanced stage. This study evaluates the performance of the Convolutional Neural Network (CNN) and Multi-Layer Perceptron (MLP) in classifying diabetes using the NHANES dataset (2,278 samples; 21 positive for diabetes). The models were tested with k-fold cross-validation using the metrics accuracy, precision, recall, F1-Score, and ROC-AUC. Results show high accuracy and precision (0.99), an average recall of 0.67, and an F1-Score of 0.75. A paired t-test indicates that CNN is superior in some metrics with a p-value of 0.374, though the ROC-AUC difference is not significant. CNNs can capture complex patterns in health features such as glucose, BMI, and age, whereas MLPs remain reliable as a baseline. In conclusion, both CNN and MLP have the potential to be used for tabular data-based diabetes classification, with CNN showing a tendency to be more effective in detecting non-linear patterns in the imbalanced dataset
Arsitektur Hibrida IndoBERTweet - Convolutional Neural Network (CNN) untuk Klasifikasi Ujaran Kebencian Berbahasa Gaul di Media Sosial
No full textDetecting hate speech on Indonesian social media is challenging due to slang, abbreviations, and informal expressions that hinder automated text understanding. Traditional machine learning approaches often fail to capture contextual meaning effectively. This study aims to develop a hate speech detection system for Indonesian slang by evaluating contextual embedding IndoBERTweet combined with a Convolutional Neural Network (CNN) architecture. The research compares the performance of CNN and BiLSTM models using IndoBERTweet and FastText embeddings. A dataset of 1,477 labeled tweets categorized as Hate Speech, Abusive, or Non-Hate Speech was used. Evaluation metrics employed in this study consist of accuracy, precision, recall, F1 score, and AUC ROC. The results show that the IndoBERTweet + CNN model achieves the best performance, with 91.2% accuracy and a 91.1% F1-score, significantly outperforming FastText-based models. IndoBERTweet’s contextual embedding proves effective in handling the linguistic complexity and implicit meanings commonly found in Indonesian slang. These findings highlight the model’s strong capability for robust hate speech detection and open opportunities for its adoption as an automated content-moderation module that identifies and filters toxic narratives on social media platforms
Evaluasi Performa YOLOv12 untuk Deteksi Plat Nomor Kendaraan Real-Time pada Citra Closed-Circuit Television (CCTV)
Full text linkLicense plate detection is a crucial component of intelligent transportation systems. Deep learning methods still face limitations in detecting small-sized plates under low-light conditions and complex backgrounds. This study evaluates YOLOv12\u27s performance for license plate detection in CCTV imagery containing small objects with great visual detail. Unlike YOLOv11, which focuses on detection efficiency for larger objects, YOLOv12 integrates attention mechanisms to enhance sensitivity to fine-grained spatial features. Model evaluation was conducted using precision, recall, and mean average precision (mAP) metrics on traffic image datasets with daytime and nighttime lighting conditions and CCTV viewing angles. Results show the model achieves [email protected] of 87.2% and precision of 89.5%, comparable to previous YOLO-based studies. However, performance drops to 47.9% at [email protected]:0.95, indicating limitations in bounding-box localization precision under visually complex conditions. This study highlights opportunities for future improvement through dataset expansion and parameter optimization for training
Analisis Kekuatan dan Potensi Resonansi pada Pondasi Diesel Generator Menggunakan Metode Elemen Hingga
Full text linkThe ship’s electrical system heavily relies on a safe and reliable diesel generator foundation. However, many foundation designs currently used in the industry have not been comprehensively validated against classification standards, nor analyzed for resonance potential and fatigue life. This study aims to develop a diesel generator foundation design that complies with ship structural safety standards based on Lloyd’s Register (LR) regulations. The methods used include 3D modeling, structural and resonance analysis using the Finite Element Method (FEM), and validation through manual calculations. The results show that the initial design does not meet safety standards, while the LR-based design yields a maximum stress of 164.01 MPa and a maximum deformation of 0.99981 mm, with a fatigue life reaching 28 years. Both designs indicate no potential for resonance. This study recommends the LR-based foundation design as the optimal solution to ensure structural integrity, long service life, and compliance with maritime safety regulations.Sebagai negara kepulauan, Indonesia membutuhkan perkembangan industri perkapalan yang signifikan. Guna mencapai fungsi dan misi kapal secara optimal, dibutuhkan sistem kelistrikan yang memadai untuk menunjang sistem pengoperasian, terutama pondasi diesel generator. Penelitian ini dilakukan dengan membandingkan desain awal dengan desain standar Lloyd\u27s Register (LR) menggunakan metode Finite Element Method (FEM) dan perhitungan manual untuk menganalisis tegangan, resonansi, dan umur kelelahan. Hasil analisis menunjukkan bahwa hanya desain LR yang memenuhi faktor keamanan, dengan tegangan maksimum 164.01 MPa dan deformasi maksimum 0.99981 mm. Kedua desain tidak berpotensi terjadi resonansi. Umur kelelahan desain LR mencapai 28 tahun, melebihi standar minimum 20 tahun. Penelitian ini merekomendasikan desain pondasi generator berbasis standar LR sebagai solusi paling optimal, sehingga dapat memastikan keamanan dan tetap mengikuti standar maritim
Digital Smart Marketing: Membangun UMKM Tangguh di Era Ekonomi Digital 5.0
Full text link Micro, Small, and Medium Enterprises (MSMEs) play a vital role in strengthening the local economy of Sunyalangu Village, Karanglewas District, Banyumas Regency. However, most MSMEs still face limitations in utilizing digital technology, particularly in marketing, business administration, and transaction management. This community service activity, titled "Building Resilient MSMEs in the Digital Economy 5.0 Era," aims to improve digital literacy and technology-based business management skills among MSMEs. This program was implemented using a gradual training, mentoring, and counseling approach from October to November 2025. A total of 30 MSMEs participated in the activities, which included digital literacy counseling, digital marketing training through social media and marketplaces, and assistance in creating online store accounts and digital-based administration systems. The results of the activity showed that most participants successfully established active marketplace accounts and began implementing online marketing strategies, which have resulted in an expanded market reach and increased business operational efficiency. However, limited internet infrastructure and the availability of digital devices remain obstacles to optimizing technology utilization. Overall, this activity has had a positive impact on changing the business behavior of MSMEs and strengthening their business independence and resilience in the face of the digital economy\u27s development.
Pengaruh Post Welding Heat Treatment (PWHT) Pada Friction Stir Welding Pada Material Aluminium AA5052/6061 Terhadap Sifat Mekanik dan Struktur Mikro
No full textThe development of technology in the maritime sector is very rapid in the manufacture of aluminum ships. Friction Stir Welding (FSW) is a green technology in the metal joining process. Post Welding Heat Treatment (PWHT) is one way to increase the loss of strength of the material connection after the FSW process. The purpose of this study was to determine the effect of FSW results before and after PWHT. Aluminum alloy (AA) series 5052 and 6061 were used as research materials because they are resistant to seawater. FSW parameters were spindle speed 1500 rpm (clockwise), travel speed 30 mm/min, tool depth 0.2 mm, and tool tilt 1o. PWHT parameters were heating the FSW material in a furnace to a temperature of 550°C (10, 30 minutes), water quenching, and artificial aging to a temperature of 180°C for 7 hours. The material was characterized by tensile testing, hardness, and micrography. The highest tensile strength and hardness values were 216 MPa and 68 HV, the smallest and uniform grain size in the weld zone and base metal areas were obtained at PWHT parameters of 30 minutes
Rancang Bangun Jig Welding Modifikasi Pneumatic Clamp Untuk Pengelasan Rear Tube Sepeda Motor Listrik
Full text linkWelding of electric motorcycle rear tube frames at PT Ganding Toolsindo still uses manual clamp welding jigs, which cause problems with long part setup times and less precise weld joints. This study aims to design a more efficient welding jig, which can increase productivity and welding quality by modifying the pneumatic clamp on the welding jig used. The research methods used include literature studies, field observations, data analysis and collection, welding jig design with pneumatic clamps, manufacturing, and testing of welding jigs. Welding with manual clamp welding jigs requires a part setup time of 40 seconds. While the part setup in welding with pneumatic clamp modified welding jigs is 19 seconds. The reduction in part setup time by using pneumatic clamp modified welding jigs is 21 seconds or 52%. These results prove that the use of pneumatic clamp modified welding jigs can increase welding efficiency and produce more precise weld joints.Pengelasan rear tube frame sepeda motor listrik di PT Ganding Toolsindo masih menggunkan jig welding manual clamp yang menimbulkan permasalahan lamanya waktu setup part dan hasil sambungan las kurang presisi. penelitian ini bertujuan untuk merancang jig welding yang lebih efisien, yang dapat meningkatkan produktivitas dan kualitas pengelasan dengan melakukan modifikasi pneumatic clamp pada jig welding yang digunakan. Metode penelitian yang digunakan antara lain studi literature, observasi lapangan, analisis dan pengumpulan data, desain jig welding dengan pneumatic clamp, pembuatan dan uji coba jig welding. Pada pengelasan dengan jig welding manual clamp, waktu yang dibutuhkan untuk setup part selama 40 detik. Sedangkan setup part pada pengelasan dengan jig welding modifikasi pneumatic clamp adalah sebesar 19 detik. Hal ini menunjukkan bahwa penggunaan jig welding modifikasi pneumatic clamp dapat menggurangi waktu proses sebesar 21 detik atau sebesar 52%
Pemanfaatan Algoritma Random Forest Regression dalam Memprediksi Kepuasan Mahasiswa Terhadap Dosen
Full text linkStudent satisfaction with lecturers is a key indicator in assessing the quality of higher education. However, commonly used evaluation approaches remain largely descriptive and subjective, making them less effective in supporting sustainable quality improvement. Moreover, the comprehensive use of lecturer competency indicators in predictive models is still limited. This study addresses the gap by developing a student satisfaction prediction model using the Random Forest Regression algorithm, optimized through grid search and feature selection using the Recursive Feature Elimination (RFE) method combined with 5-fold cross-validation. Data were collected from the EDOM system of Politeknik Negeri Cilacap, involving 24 indicators based on national lecturer competency standards, and analyzed using R software. The best model was achieved with parameters mtry = 1 and ntree = 300, yielding RMSE = 0.0222, MAE = 0.0118, and R² = 0.9959. The three most influential indicators identified were structured assignments, diversity of teaching methods, and punctuality. These findings are expected to inform policies for improving the quality of higher education
Rancang Bangun dan Uji Kinerja Mesin Oil Skimmer Disk Type Berdasarkan Volume dan Kadar Ph Pada Cairan Pendingin Mesin CNC Milling Hurco Seri Vm10
Full text linkModern machining processes heavily rely on liquid coolants to extend the life of cutting tools; however, water-soluble coolants are susceptible to oil contamination, which degrades their quality. This research aims to design the form and details of a disc-type oil skimmer machine, perform the necessary machine element calculations, and test the results. The design process utilizes the VDI 2222 systematic approach with the aid of Autodesk Inventor 2025 software, supported by literature and field studies. The calculation results indicate a required motor power of 0.042 HP, a minimum shaft diameter of 7.2 mm, and a dynamic bearing load of 5.868 N. The frame strength analysis also ensures safety, with a maximum stress of 15.19 N/mm², well below the allowable limit of 79,2 N/mm². In conclusion, this design yields a viable and safe oil skimmer for implementation on the Hurco VM10 series CNC Milling machine, complete with detailed working drawings and an optimal speed of 80 rpm for separating oil from the coolant
Support Vector Machine (SVM) - Based Optimization of Leukemia Cell Image Classification
Full text linkLeukemia is a type of blood cancer characterized by the uncontrolled proliferation of abnormal white blood cells that originate from the bone marrow. Early detection of leukemia poses a significant challenge in the medical field, as the conventional diagnostic process still relies on manual microscopic observation by hematologists, which is time-consuming and prone to subjective errors. This study aims to analyze the potential of the Support Vector Machine (SVM) algorithm in optimizing the classification of leukemia cell images based on morphological and texture features extracted from microscopic images. The test results show that the SVM model with the RBF kernel provides the best performance with an accuracy of 96.4%, a precision of 95.8%, a recall of 96.1%, and an F1-score of 96.0%, surpassing the results of linear and polynomial kernels. The analysis shows that the use of a combination of shape and texture features has a significant effect on improving classification accuracy