Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
A Comparison of MobileNetV2 and VGG16 Architectures with Transfer Learning for Multi-Class Image-Based Waste Classification
Get PDFEffective waste management represents a global challenge with significant environmental and public health impacts. Despite existing waste classification systems achieving high accuracy rates, a critical research gap exists in determining optimal CNN architectures for real-world deployment constraints, particularly regarding computational efficiency versus classification accuracy trade-offs. We compared two Convolutional Neural Network (CNN) architectures MobileNetV2 and VGG16 for classifying ten types of waste using image-based analysis. Using transfer learning approach, both models were modified for waste classification tasks by adding custom layers to pre-trained models. The dataset contained 19,762 images balanced to 9,440 samples through under-sampling techniques and enhanced with data augmentation to increase variation. Results demonstrated that MobileNetV2 achieved 95.6% test accuracy with precision 0.93, recall 0.93, and F1-score 0.93, significantly outperforming VGG16\u27s 89.13% accuracy with precision 0.91, recall 0.90, and F1-score 0.90. Beyond superior accuracy, MobileNetV2 also demonstrated higher computational efficiency with 350ms/step training time compared to VGG16\u27s 700ms/step, and more consistent performance across all waste categories
Performance Comparison of Embeddings and Keyword Selection Methods in Enterprise Document
Get PDFKeyword extraction is widely used in domains such as social media and e-commerce, but its application for enterprise document retrieval remains limited. Most organizations still depend on structured systems or rule-based approaches for indexing, which often lack semantic understanding and scalability. While several techniques like TextRank and RAKE have been explored, few studies assess their effectiveness on operational document retrieval in institutional settings, revealing a research gap. This study investigates the use of KeyBERT to extract keywords from university documents, including SOPs, manuals, and guidelines. KeyBERT leverages transformer-based embeddings to generate semantically relevant keywords and is chosen for its ease of use, model flexibility, and no need for labeled data. Additionally, it supports diversification strategies such as Maximum Marginal Relevance (MMR) and MaxSum to reduce redundancy and enhance keyword variety. We evaluate six embedding models combined with three keyword selection methods: Cosine similarity, MMR, and MaxSum. The best F1 score of 0.78 is achieved using Cosine with the paraphrase-MiniLM-L3-v2 model, along with an average extraction time of 184.02 seconds. These findings highlight the effectiveness of combining lightweight embeddings with strategic keyword selection for enterprise-scale document indexing
Cybersecurity Awareness in Government Institutions: A Systematic Review of Behavioral Strategies and Policy Readiness
Get PDFThis study presents a systematic literature review of cybersecurity awareness in government institutions, focusing on how public agencies perceive, implement, and sustain awareness initiatives. It critically examines behavioral strategies, training methods, and policy frameworks that influence employee engagement and institutional readiness. Findings highlight the fragmentation of awareness efforts, gaps in policy coordination, and the need for behaviorally informed, role-specific interventions to build cyber resilience in the public sector. Recommendations include strengthening institutional infrastructure, leadership engagement, and adaptive learning strategies to enhance cybersecurity awareness and long-term resilience
Detection and Localization of Brain Tumors on MRI Images Using the YOLO Algorithm
Get PDFThis study addresses the critical need for early and accurate brain tumor diagnosis on MRI images by comparing five versions of the YOLO algorithm (YOLOv5, YOLOv7, YOLOv8, YOLOv9, and YOLOv12) with consistent parameters. Utilizing a pre-annotated Kaggle MRI brain dataset, the research meticulously verified annotations and employed data augmentation (flipping, rotation, blurring, noise) to expand the dataset from 801 to approximately 1362 images, enhancing model generalization and robustness. Models were trained and evaluated on metrics including precision, recall, [email protected], [email protected]:0.95, and inference time. YOLOv12 demonstrated superior overall performance, achieving the highest recall (97.32%), [email protected] (92.2%), and [email protected]:0.95 (76.57%), establishing its robustness for accurate detection and object localization. While YOLOv7 achieved the highest precision (96.89%) and excellent inference speed, its overall mAP and recall were surpassed by other iterations. YOLOv9 and YOLOv8 also showed strong competitive performance, indicating significant advancements in the newer YOLO generations. The findings confirm the efficacy of the YOLO algorithm for brain tumor detection and localization in MRI images, with YOLOv12 proving to be the most effective variant in this comparative analysis
Personal Protective Equipment Completeness Monitoring System Using YOLO-Based Computer Vision
Get PDFWorkplace safety in the construction sector remains a critical concern, primarily due to low compliance with Personal Protective Equipment (PPE) standards. To address this, this study develops and evaluates a real-time PPE monitoring system, conducting a comparative analysis of two state-of-the-art object detection models: YOLOv8s and YOLOv11s. The system is designed to detect three essential PPE items: helmets, masks, and vests, and both models were trained on a custom dataset of 9,202 augmented images over 200 epochs. The final evaluation on an unseen test set revealed highly competitive performance. While YOLOv8s achieved a marginally higher [email protected] (90.8%), YOLOv11s demonstrated superior precision (92.0%) and better performance on the stricter [email protected]:0.95 metric (54.4%). Based on this nuanced trade-off and its significantly higher computational efficiency (15% fewer parameters), YOLOv11s was selected as the optimal model. The chosen model achieved a real-time inference speed of approximately 112 FPS. A functional web-based prototype was developed using Flask to demonstrate the system\u27s practical application. These findings confirm that YOLOv11s offers a more balanced and efficient solution for automating PPE compliance monitoring and highlight that a holistic evaluation beyond a single metric is crucial for deploying robust computer vision systems in real-world safety applications.Workplace safety in the construction sector remains a critical concern, primarily due to low compliance with Personal Protective Equipment (PPE) standards. To address this, this study develops and evaluates a real-time PPE monitoring system, conducting a comparative analysis of two state-of-the-art object detection models: YOLOv8s and YOLOv11s. The system is designed to detect three essential PPE items: helmets, masks, and vests, and both models were trained on a custom dataset of 9,202 augmented images over 200 epochs. The final evaluation on an unseen test set revealed highly competitive performance. While YOLOv8s achieved a marginally higher [email protected] (90.8%), YOLOv11s demonstrated superior precision (92.0%) and better performance on the stricter [email protected]:0.95 metric (54.4%). Based on this nuanced trade-off and its significantly higher computational efficiency (15% fewer parameters), YOLOv11s was selected as the optimal model. The chosen model achieved a real-time inference speed of approximately 112 FPS. A functional web-based prototype was developed using Flask to demonstrate the system\u27s practical application. These findings confirm that YOLOv11s offers a more balanced and efficient solution for automating PPE compliance monitoring and highlight that a holistic evaluation beyond a single metric is crucial for deploying robust computer vision systems in real-world safety applications
Information Cascades in Professional Networks: A Graph-Based Study of LinkedIn Post Engagement
Get PDFInformation cascades in professional networks represent a critical mechanism for knowledge transfer and career development, yet their dynamics remain poorly understood. This study presents a comprehensive empirical analysis of information cascades in LinkedIn professional networks, focusing on computer science professionals and academic-industry knowledge transfer. We analysed 50,000 CS professionals, 500,000 connections, and 100,000 technical posts over 12 months using a Modified Independent Cascade Model that incorporates professional context factors. Our analysis reveals that hybrid professionals, representing only 25% of the network, account for 52% of inter-cluster connections and achieve 2.8× higher cross-domain transfer rates. Educational content demonstrates superior cross-domain appeal (0.47) compared to research papers (0.23), with optimal posting windows between 10 AM-12 PM achieving 23% higher cross-domain engagement. Bridge users in academic-industry transitions show significantly higher transfer effectiveness (Cohen\u27s d = 1.47, p < 0.001). These findings provide evidence-based strategies for optimising professional networking and knowledge dissemination across academic and industry domain
The Influence of Workplace Environment and Work Motivation on Job Performance: The Mediating Role of Employee Commitment Among Generation Z Workers
Get PDFThis study examines the influence of workplace environment and work motivation on job performance among Generation Z employees within Indonesia’s food and beverage (F&B) sector, with a particular focus on the mediating role of employee commitment. Employing a quantitative research approach, data were collected through structured online surveys and analyzed using structural equation modeling. The findings reveal that both workplace environment and work motivation significantly enhance employee commitment and directly improve job performance. Furthermore, employee commitment was found to partially mediate the relationships between workplace environment, motivation, and job performance, indicating its crucial role in sustaining organizational outcomes despite Generation Z’s typically lower organizational loyalty. The research highlights the importance of fostering flexible, supportive, and inclusive work environments tailored to the unique preferences of Generation Z employees. Additionally, intrinsic motivators such as personal growth opportunities and recognition were identified as key drivers of commitment and performance. This study contributes to the literature by integrating Herzberg’s Two-Factor Theory, the Job Demands-Resources Model, and Social Exchange Theory to provide a comprehensive framework for managing Generation Z in high-turnover service industries. Practical recommendations are offered for human resource strategies, while suggestions for future research call for exploration across broader industrial and cultural settings
Interaction Use Paylater Regarding Student Decisions with Lifestyle as Variables Moderation in the Perspective of Islamic Economics : A Case Study of FEBI Students at UIN North Sumatra
Get PDFThe advancement of financial technology has given rise to Paylater services, which are increasingly favored by university students, particularly for fulfilling their consumption needs directly. However, this convenience has the potential to trigger consumer behavior that conflicts with it. Based on this phenomenon, this study formulates the following question: Does the use of a Paylater perspective influence students\u27 consumption decisions, and does lifestyle moderate this relationship? This study aims to analyze the effect of Paylater usage on student decisions and to examine the role of lifestyle as a moderating variable. The method employed is a quantitative approach using multiple regression and Moderated Regression Analysis (MRA). A sample of 98 students from the Faculty of Islamic Economics and Business at UIN North Sumatra was obtained through purposive sampling. The results indicate that Paylater usage has a significant effect on students\u27 consumption decisions. Furthermore, lifestyle is proven to moderate this relationship, where a consumptive lifestyle strengthens the influence of pay later on consumption decisions. These findings indicate a shift in student consumption patterns toward instant and emotional behavior. Therefore, financial literacy education based on Sharia principles and promotion is necessary to address the penetration of digital financial services
Pengaruh Good Corporate Governance terhadap Kinerja Keuangan Pada Perusahaan Transportasi dan Logistik
No full textThe effect of good corporate governance on financial performance is the focus of this study. by involving control variables, namely Company Size and covid-19 dummy variables. The study’s population consist of Transportation and Logistics companies listed on the Indonesia Stock Exchange for the period 2021-2023, there are 37 companies. 23 companis ware sampled in this study based on criteria. This research methodology uses multiple linear regression analysis. This data processing utilizes the SPSS Ver26 application. Considering the findings of the analysis, this research indicates that managerial ownership, institutional ownership and audit committee have a positive effect on financial performance. but the independent board of commissioners and the audit committee have no effect on financial performance.Pengaruh good corporate governance terhadap kinerja keuangan menjadi fokus penelitian ini, dengan melibatkan variabel kontrol yaitu Ukuran Perusahaan dan variabel dummy covid-19. Populasi penelitian ini adalah perusahaan Transportasi dan Logistik yang terdaftar di Bursa Efek Indonesia periode 2021-2023 sebanyak 37 perusahaan. Sebanyak 23 perusahaan dijadikan sampel dalam penelitian ini berdasarkan kriteria yang telah ditetapkan. Metodologi penelitian ini menggunakan analisis regresi linier berganda. Pengolahan data ini memanfaatkan aplikasi SPSS Ver26. Berdasarkan hasil analisis, penelitian ini menunjukkan bahwa kepemilikan manajerial, kepemilikan institusional dan komite audit berpengaruh positif terhadap kinerja keuangan. Namun dewan komisaris independen dan komite audit tidak berpengaruh terhadap kinerja keuanga
Capability Level Assessment of IT Governance in the SIAP KOJA Application Using the COBIT 2019 Framework
No full textThis study assesses the IT governance capability of the SIAP KOJA application at the Jambi City Department of Communication and Informatics (Diskominfo). SIAP KOJA was introduced to strengthen attendance discipline, transparency, and accountability through geofencing and biometric features. Using the COBIT 2019 framework, the assessment aligns IT processes with institutional objectives and focuses on two key processes: APO11 (Manage Quality) and BAI05 (Manage Organizational Change Enablement). Data were collected through a literature review, interviews, observations, and a structured questionnaire based on the COBIT 2019 Process Assessment Model. The sample comprised five personnel from Diskominfo’s Informatics Applications Division, purposively selected for their direct involvement in planning, development, and operations. Results indicate Level 3 (Defined) capability for both APO11 and BAI05 with standards documented. At Level 4, APO11 reached 75.56% (Largely Achieved) and BAI05 reached 76.00% (Largely Achieved). Because these fall below the ≥85% “Fully Achieved” threshold, progression was halted, and the capability level remains Level 3. Limitations in structured measurement and continuous monitoring contribute to a two-level gap from the Level 5 (Optimizing) target. The study recommends formalizing a quality management system with service-level agreements and performance indicators; strengthening outcome-based change management through compliance audits and systematic user feedback; and institutionalizing lessons learned. These improvements are essential for enhancing governance capability, ensuring system reliability, and supporting successful digital transformation in local government