Nha Trang University

This study provides new insights by quantifying the role of human capital in driving firm-level innovation, focusing on the transition and developing economy of Vietnam. Our research broadens the perspective on diverse categories of human capital and their influence on firms' innovation. The empirical analysis utilizes a rich dataset constructed from the two most recent waves of World Bank Enterprise Surveys, covering 2,005 observations for both domestic and foreign-invested firms in Vietnam. The estimation results show that basic human capital has a limited influence on firm-level innovation, while firm-invested and scarce human capital play a more substantial role. Additionally, we find that firm size, R&D activity, exporting, and informal competition facilitate firm-level innovation, whereas foreign ownership and firm age exert the opposite impact. Our disaggregated analyses across firm sizes and sectors suggest divergent effects of human capital, offering deeper insights for policymakers and managers seeking targeted strategies to foster innovation.

Accurate prediction of Type 2 Diabetes is essential for effective prevention and intervention strategies. This paper proposes a hybrid model that integrates an improved Louvain community detection algorithm with a domain-specific diabetes ontology to enhance prediction performance and interpretability. Three real-world datasets (Pima Indians, Diabetes 130-US Hospitals, and NHANES) were used for evaluation. Patients were grouped into clinically meaningful clusters using the enhanced Louvain method, followed by semantic reasoning over an ontology constructed from expert-defined medical rules and clinical features. Experimental results show that the proposed model outperforms conventional clustering and classification techniques across multiple metrics, achieving up to 97.56% accuracy in cross-validation. The use of ontological knowledge not only increases transparency in prediction outcomes but also supports semantic querying and dynamic model adaptation. This method demonstrates significant potential for real-world deployment in clinical decision support systems.

This investigation presents the nonlinear vibration analysis of sandwich double curved shallow shell on elastic foundations in the thermal environment and subjected to electric, magnetic, and mechanical fields. The sandwich double curved shallow shell comprises a particle- and fiber-reinforced composite (PFRC) core and magneto-electro-elastic (MEE) face sheets. Analytical expressions of material properties for the particle- and fiber-reinforced composite are derived from the volume proportions of the polyester matrix, particles, and fibers. The coupling motion equations are derived using Reddy’s first-order shear deformation shell theory and the von Kármán geometric nonlinearity. Natural frequency, phase plane trajectory, and dynamic response of the sandwich double curved shallow shell are found using the Galerkin and Runge–Kutta methods. Numerical findings evaluate the effect of various factors encompassing dimensional attributes, fiber and particle volume fraction, temperature rise, electric and magnetic potentials, and elastic foundations coefficients on vibration characteristics of the sandwich double curved shallow shell.

Small-scale fisheries are crucial for supporting the welfare of coastal communities. Nonetheless, in Vietnam prolonged overexploitation and inadequate management have led small-scale fisheries into an uncertain future, leaving fishing households vulnerable to poverty and food insecurity. This study examines the role of small-scale fisheries in Vietnam in promoting food security and alleviating poverty within fishing households. Utilizing latent profile analysis, we categorize fishing households based on dimensions of poverty and food insecurity as well as explore the potential of fisheries management measures in eradicating poverty and improving food security. Our findings reveal that, small-scale fisheries in Vietnam have significantly contributed to the well-being of fishing households, enhancing both income and food security. However, we identify two distinct groups of fishers. One group, representing 65 percent of households in our sample, is characterized by higher incomes and greater food security, is denoted in the study as “protected households”. The second group, comprising 35 percent of our sample, faces challenges in both dimensions, and is denoted as “vulnerable households”. Protected households are more likely to be located in areas where access limitations are enforced, often accompanied by livelihood enhancement opportunities. These results suggest that future policies for small-scale fisheries could benefit from developing synergies among various interventions targeting the conservation of fisheries resources, poverty alleviation, and food security.

Causal inference is a cornerstone of biomedical reasoning, yet current methods often struggle to integrate domain knowledge, handle uncertainty, and provide interpretable results. In this paper, we propose a novel framework - Causal Reasoning over Ontology-Enriched Graphs (CREOG) - that combines biomedical ontologies with causal discovery and knowledge graph embeddings to facilitate explainable causal analysis. we construct a diabetes ontology and derive a small causal knowledge graph (CKG) and an enlarged SNOMED-anchored diabetes CKG with 21,542 entities and 107,710 causal or associative edges. We additionally validate on a breast cancer ontology (BreastCancer.owl) to assess cross-domain generalizability. CREOG combines association rule mining (Apriori) with multiple causal discovery algorithms (PC, NOTEARS, FCI, GES/GIES) to generate weighted causal triples with textual provenance. To enable robust causal link prediction, we introduce CausE-S, a causal embedding model that leverages edge weighting and structural similarity to prioritize high-confidence relations. Experimental results show that CausE-S significantly outperforms traditional embedding baselines such as TransE, ComplEx, GraIL, and R-GCN on both KGs across multiple evaluation metrics, including MRR, Hits@K, Precision, Recall, and F1-score. Additional comparisons to FCI and GES/GIES confirm that CREOG recovers a large fraction of baseline causal edges while proposing clinically plausible new relations. The proposed framework also supports explainable reasoning through annotated causal links and confidence-aware inference. Our work demonstrates the effectiveness of combining ontological semantics with causal embeddings for interpretable biomedical knowledge discovery.