Guruh Fajar Shidik

Early detection of plant leaf diseases is critical for minimizing crop losses and supporting precision agriculture. While Convolutional Neural Networks (CNNs) have demonstrated high accuracy in image-based diagnosis, conventional architectures may not optimally balance spatial localization and channel-wise feature refinement, particularly in multi-crop classification settings. This study proposes a redundancy-aware dual-attention architecture, termed ATSA-DenseNet, which integrates the spatial branch of the Convolutional Block Attention Module (CBAM-Spatial) with Efficient Channel Attention (ECA) within a DenseNet121 backbone. Unlike prior dual-attention frameworks that retain full CBAM and introduce channel-level redundancy, the proposed design isolates complementary spatial and channel mechanisms to improve representational efficiency without increasing computational complexity. The framework is evaluated on controlled multi-crop PlantVillage-derived datasets comprising tomato, potato, pepper, and maize. Across both 3-crop and 4-crop configuration, ATSA-DenseNet consistently outperforms baseline DenseNet121 and single-attention variants, achieving 99.94% accuracy and 0.9994 macro-F1 on the 4-crop setting while maintaining a lightweight footprint (6.96M parameters, 2.87G FLOPs). Grad-CAM visualizations indicate improved localization of disease-relevant regions compared to the baseline. While results are obtained under controlled imaging conditions, the findings demonstrate that redundancy-aware dual-attention enhances feature discrimination efficiency in multi-class agricultural classification tasks. Future work will extend validation to real-field datasets with natural variability.

Digital transformation constantly changes the work practices and employee experiences of contemporary work environments. Studies have documented the adverse impact of cutting-edge technologies, such as artificial intelligence, algorithmic systems, and digital platforms on employees’ well-being. However, such findings remain fragmented across technologies, disciplines, and well-being constructs, limiting a coherent understanding of how digitally transformed work conditions affect employees. This study systematically reviews the literature on digital transformation and employee well-being to clarify the conditions, mechanisms, and outcomes that are most consistently identified in prior research. Using the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) method, this study identified, screened, and selected 57 peer-reviewed articles (2014–2025) on digital transformation and employee well-being. The Gioia inductive analytical approach was used to synthesise the reviewed studies and to develop higher-order conceptual dimensions. Five interrelated aggregate dimensions were identified: digital transformation conditions, digital resources and demands, mediating processes, contextual factors, and employee well-being outcomes. The findings indicate that employee well-being depends significantly on how digitally intensified demands and available resources are configured, interpreted, and mediated within specific organisational contexts. This review highlights the need for more temporally sensitive, context-specific, and resource-oriented research, and proposes an integrative framework to guide future research and support the design of healthier and sustainable digital workplaces.

This study applies the K-Means clustering algorithm to profile market vendors in Semarang City based on their Ability To Pay (ATP) and Willingness To Pay (WTP) BPJS Health insurance contributions, addressing a notable decline in participation among Non-Salaried Workers (PBPU) in Indonesia's National Health Insurance (JKN-KIS). Utilizing quantitative data from 95 Bulu Market vendors, ATP and WTP were estimated via household income and expenditure analysis through linear regression models performed in R-Commander (p < 0.05). Clustering with k=4 yielded optimal segmentation (Davies-Bouldin Index = 0.079), identifying four distinct groups: Opportunistic (low ATP/WTP), Rational (high ATP/moderate-low WTP), Price-Sensitive (financially vulnerable with low ATP/WTP), and Prospective (high ATP and WTP). This novel integration of ATP and WTP in Indonesia's informal sector provides data-driven insights allowing BPJS policy makers to tailor intervention strategies, improving premium compliance within the JKN system. The methodology offers a replicable framework for health insurance segmentation in comparable middle-income Southeast Asian settings, supporting Universal Health Coverage goals.

Tomato is one of the widely available horticultural products and holds significant economic value in Indonesia. However, its productivity is often disrupted by various leaf diseases. This study aims to compare the performance of three CNN architectures—DenseNet121, Xception, and MobileNetV2—in classifying tomato leaf diseases. The dataset used consists of 10,000 balanced images across ten classes: Bacterial Spot, Septoria Leaf Spot, Early Blight, Late Blight, Mosaic Virus, Yellow Leaf Curl Virus, Leaf Mold, Target Spot, Spider Mites Two-Spotted Spider Mite, and Healthy. All images were resized to 224x224 pixels and divided into 80% training data and 20% test data. Augmentation techniques were applied to balance the data across classes. Experimental results show that the Xception architecture outperforms the other models, achieving an accuracy of 98.79%, with a precision of 98.80%, recall of 98.79%, and an F1-Score of 98.78%. These findings indicate that the Xception model is highly effective for plant disease classification and is suitable for implementation in environments with limited resources.

As global demand for cloud computing rises, green computing has become essential. Federated Edge Cloud (FEC) offers improved energy efficiency compared to traditional infrastructures, yet managing distributed energy consumption remains a challenge. This research introduces an Unsupervised Cluster Reinforcement Q-Learning method in FEC (UCRL-FEC), which integrates Fuzzy C-Means (FCM) or K-Means clustering to identify migratable Virtual Machines (VMs) from overloaded hosts. The method enhances energy efficiency and workload balance by incorporating a modified reward function in Q-Learning. Experimental evaluations demonstrate that UCRL-FEC reduces energy consumption (EC) up to 1.07%, contributing to lower operational costs and a reduced carbon footprint, which is critical for large-scale cloud environments. In terms of Service Level Agreement Time per Active Host (SLATAH), UCRL-FEC achieves an improvement up to 1.56% over the baseline method, demonstrating enhanced efficiency in managing active host resources. Additionally, system stability improves with up to 9.68% reduction in Performance Degradation due to Migration (SLA-PDM), effectively minimizing service disruptions and ensuring efficient workload management. Furthermore, the method reduces overall Service Level Agreement Violations (SLAV) up to 6.06%, indicating enhanced service reliability and optimized resource allocation. A Friedman test confirms statistically significant improvements in energy efficiency, workload distribution, and system stability over baseline methods. These advancements prevent resource overutilization, enhance workload management, and extend hardware lifespan, fostering sustainable cloud operations. UCRL-FEC balances energy efficiency, performance, and scalability through dynamic resource optimization, making it a viable solution for intelligent VM management in modern cloud-edge infrastructures.