Journal of Hunan University Natural Sciences

In decentralized learning scenarios such as federated learning (FL), client data are often highly heterogeneous, leading to non-IID distributions that degrade global model performance, slow convergence, and limit personalization. Conventional FL aggregation schemes inadequately capture client-specific structure and struggle to generalize across diverse data regimes. In this work, we propose a personalized federated learning framework that integrates Fisher Discriminant Analysis (FDA), meta-learning, and convolutional neural networks (CNNs) to jointly enhance class separability, adaptability, and representation quality. FDA is employed to project client-side features into an optimal class-separable latent space, enabling more discriminative modeling of local task characteristics under non-IID conditions.
On top of this representation, a meta-learning strategy is used to learn initialization parameters that support fast adaptation, allowing each client to efficiently fine-tune the model to its own data while preserving global knowledge. A CNN backbone is adopted to extract hierarchical feature representations, providing both fine-grained local patterns and robust global semantic embeddings suitable for heterogeneous federated environments. We conduct extensive
experiments on CIFAR-10, CIFAR-100, and EMNIST-Writers under non-IID data partitions generated using a Dirichlet concentration parameter α = 0.5. The proposed method achieves 79.6% test accuracy on CIFAR-10, 45.7% on CIFAR-100, and 85.5% on EMNIST-Writers, consistently outperforming strong baseline FL methods in terms of adaptability, generalization, and personalization quality. The results demonstrate that combining class-separable latent spaces with meta-adaptation improves both global coherence and client-specific performance while preserving data privacy and offering good scalability. Future work will extend the framework to multi-modal and sequential data and investigate its deployment in cross-device and cross-silo FL settings.

Keywords: federated learning; personalized models; Fisher Discriminant Analysis; meta-learning; non-IID data; data privacy.

DOI https://doi.org/10.55463/issn.1674-2974.52.11.1

For Muslims, the Qur’an functions as a primary and enduring source of law that governs all dimensions of life, including economic exchange in increasingly digital markets. This article examines how Qur’anic principles and classical fiqh doctrines can be applied to contemporary e-commerce practices in Indonesia and how such an application may support the development of the Islamic economy in the context of the Fourth Industrial Revolution (Industry 4.0). The study adopts a qualitative design and draws on a structured review of Qur’anic verses related to trade and ethics, authoritative classical works on Islamic commercial law, modern fatwas, and recent empirical and policy-oriented studies on digital trade in Indonesia. Particular attention is devoted to the fiqh analysis of online transactions through the lens of the as-salam contract and other relevant contractual forms, including agency (wakālah), brokerage (samsārah), and deferred sale (bayʿ bi-taʾjīl). The analysis suggests that e-commerce is, in principle, permissible (mubāḥ) provided that it meets the core conditions of a valid Islamic sale: clear specification of the subject matter, price, and delivery terms; the absence of ribā, excessive gharar, and maysir; and the exclusion of intrinsically prohibited goods and services. The article further argues that digital instruments such as payment gateways, electronic signatures, escrow arrangements, and certification mechanisms can enhance, rather than weaken, Sharīʿah compliance when embedded in an appropriate regulatory and supervisory framework. Overall, the study contends that a coherent Qur’an-based legal and ethical architecture for e-commerce can strengthen transactional trust, broaden the participation of Muslim consumers and entrepreneurs, and contribute to a more inclusive and resilient Islamic digital economy in Indonesia.

Keywords: e-commerce; Qur’an; Industry 4.0; Islamic economics; as-salam contract; Indonesia.

DOI https://doi.org/10.55463/issn.1674-2974.52.10.9

Peat soil presents major challenges for civil engineering applications due to its high acidity, low bearing capacity, low shear strength, and high water content. This study investigates the improvement of peat soil shear strength through a combined stabilization approach: grouting using industrial and agricultural waste materials, followed by biogrouting employing indigenous Bacillus cereus. Calcium carbide residue (CCR) and rice husk ash (RHA) were used as grouting agents to supply calcium and silica for the formation of calcium silicate hydrate (C–S–H) compounds. Subsequently, biogrouting was carried out on selected specimens using B. cereus isolated from peat soil in Cikarang, Indonesia, to induce CaCO₃ precipitation. Bacterial treatment was applied 30 days after grouting. For the grouting-only specimens, cohesion increased by 124.62% with an internal friction angle gain of 5° after 30 days; after 60 days, cohesion increased by 137.58% and friction angle by 7°; and after 90 days, cohesion reached a 148.13% increase relative to untreated peat. The incorporation of B. cereus via biogrouting further enhanced shear strength. After 30 days of biogrouting, cohesion increased by 115.73% relative to the 60-day grouting-only condition, with a 2° friction angle. After 60 days of biogrouting, cohesion improved by 151.03% compared with the 30-day biogrouting condition, with a 3° friction angle; after 112 days, cohesion rose by an additional 32.83%, with a 4° friction angle. The study provides three main contributions: the combined use of CCR and RHA as sustainable waste-based grouting materials, the application of indigenous B. cereus as a biogrouting agent, and a two-stage stabilization strategy that first reduces peat acidity through grouting and then applies biogrouting to further enhance shear strength.

Keywords: peat soil; Bacillus cereus; calcium carbide residue; rice husk ash; biogrouting; soil stabilization.

DOI https://doi.org/10.55463/issn.1674-2974.52.10.8

This study aims to examine the development of pre-service chemistry teachers’ argumentation skills during the completion of structured tasks embedded in the 5E Learning Cycle, as well as their responses to this training. Argumentation is a key component of effective chemistry learning; however, many pre-service teachers still lack confidence and the capacity to construct coherent, well-supported arguments. This research investigated the effectiveness of the 5E Learning Cycle model in strengthening argumentation skills using a quasi-experimental time-series design involving 71 students distributed across three classes in a chemistry review course. Data were collected through collaborative worksheets and individual pretest–posttest instruments and were analyzed using Toulmin’s argumentation model. The findings indicate a progressive enhancement in students’ argumentation performance, particularly in the formulation of claims, the use of data, and the articulation of warrants, backing, and rebuttals across topics such as chemical bonding, chemical equilibrium, acid–base chemistry, and organic chemistry. Nevertheless, a decrease in the quality of claims, warrants, and qualifiers was observed on the acid–base posttest, suggesting that certain content areas may require more targeted support. Overall, the results demonstrate that structured task-based activities within the 5E Learning Cycle can effectively foster argumentation skills in pre-service chemistry teachers, while also highlighting the need for refinement in topic-specific implementation. This study contributes new insights into the design of argumentation-focused training in teacher education programs aimed at strengthening the professional competence of future chemistry teachers.

Keywords: argumentation skills; Learning Cycle 5E; worksheet; Toulmin argumentation model.

DOI https://doi.org/10.55463/issn.1674-2974.52.10.7

Digital transformation in educational management has emerged as a major catalyst for systemic change across educational institutions. This study examines the role of digital transformation in driving managerial innovation, identifying implementation challenges, and assessing its influence on overall school quality. Employing a qualitative approach through an extensive literature review, the article synthesizes insights from ten reputable international journals relevant to the field of digital education transformation.
The findings indicate that digitalization has stimulated a wide range of innovations, including the use of big data for performance evaluation, interactive platform-based learning models, and predictive analytics to support evidence-based policymaking. Moreover, digital transformation enhances transparency, accountability, and stakeholder engagement throughout the educational process. Despite these benefits, several persistent challenges continue to hinder its effective implementation, such as the digital divide, resistance stemming from organizational culture, limited digital literacy among educators, and concerns regarding data privacy and security.
Nevertheless, schools that strategically integrate digital transformation demonstrate substantial improvements in managerial efficiency and educational quality. Such progress is largely influenced by adaptive digital leadership and an organizational culture that embraces continuous change. Strengthening digital infrastructure, providing ongoing professional development, and fostering cross-sector collaboration are therefore essential components for achieving sustainable digital transformation in education.
This study offers important implications for policymakers, school leaders, and education practitioners seeking to design context-appropriate and long-term strategies for effective digital transformation.

Keywords: Digital Transformation, Education Management, School Innovation, Education Quality, Digital Leadership.

DOI https://doi.org/10.55463/issn.1674-2974.52.10.6