Search Results (4,391)

Stock index forecasting is essential for financial market analysis and risk monitoring, yet it remains challenging because index price series are nonlinear, non-stationary, and driven by heterogeneous market factors. Existing methods remain limited in preserving local price patterns, capturing multi-scale local dependencies, and integrating attention-derived structures with long-range state-space representations. To address these limitations, we propose AG-SSM, an attention-guided state-space model for multi-step stock index forecasting. The model first uses variable-wise patch embedding to construct local semantic units, which are then processed by the AG-SSM architecture for temporal representation learning. Its core block integrates dual-path local attention (DPLA), S4D-based state-space feature generation, attention-guided aggregation (AGA), and gated update (GU). Specifically, DPLA combines sliding and dilated local attention to capture contiguous and sparsely distributed dependencies, while AGA reuses local attention maps to refine state-space features, thereby coupling local market structures with long-range sequential dynamics. Experiments on six stock index datasets (SSE, SZSE, SMESE, SP500, DJIA, and NIKKEI225) under one-, five-, ten-, and fifteen-step forecasting horizons show that AG-SSM achieves the lowest horizon-averaged MAPE on all six datasets while maintaining competitive performance across other metrics and individual horizons. Averaged over five independent runs, the horizon-averaged MAPE values are 1.5466%, 2.2173%, 2.2293%, 1.4373%, 1.2995%, and 1.8738% on the six datasets, respectively. Ablation studies, state-space variant comparisons, sensitivity analyses, and statistical tests further support the effectiveness and robustness of the proposed framework. Full article

Small- and medium-sized enterprises (SMEs) face significant challenges in adopting robotic solutions due to limited financial resources, insufficient technical expertise, and uncertainty regarding operational and economic outcomes. Existing automation approaches are often technologydriven and provide limited support for systematic decisionmaking. This study proposes the Agile Robotics Implementation Model (ARIM), an iterative framework integrating Lean Manufacturing, Lean Robotics, and Lean Startup principles. ARIM combines process assessment, key performance indicator (KPI)-based evaluation, and iterative experimentation within the Robotic Startup Cycle, supported by a decision-support software tool. The framework was developed using a Design Science Research (DSR) approach and validated through an industrial case study. Results demonstrate strong agreement between predicted and realized KPI values. The implemented solution achieved a 24.5% return on investment (ROI), with a payback period of approximately 2.1 years, reduced labor demand by 3900 h, and improved productivity, ergonomics, and quality. The findings indicate that ARIM supports reliable and data-driven robotics implementation in the studied SMEs; broader transferability requires validation across multiple cases. Full article

Advances in healthcare have led to substantial growth in the global population of cancer survivors, yet young adult cancer survivors (YACS) aged 18–39 years remain an underrepresented population in cancer survivorship research. This review examines the health challenges and psychosocial factors that shape nutritional status in YACS, with the aim of informing age-appropriate nutritional care. YACS face significant financial and psychosocial demands unique to this life stage, balancing education, early careers and family responsibilities. They report greater unmet nutritional and health needs than other age cohorts, in addition to concerns regarding long-term side effects, financial strain, fear of cancer recurrence and increased responsibility for managing their own care. Cancer treatment can cause nutrition impact symptoms (e.g., nausea, taste changes, fatigue, dysphagia and gastrointestinal (GI) disturbances) that may persist for years, and YACS are vulnerable to late effects such as endocrine dysfunction, cardiometabolic disease, chronic pain and osteoporosis, all of which can be influenced by diet. Poor dietary patterns have been observed in YACS, indicating low intakes of fruits, vegetables, fibre and dairy, and higher consumption of saturated fat, sodium and processed foods. This narrative review evaluated fourteen nutritional intervention studies targeting YACS aged 18–39. Most existing interventions demonstrate minimal recruitment and retention rates, small sample sizes, and a reliance on self-report methods rather than objective nutritional measures. These limitations highlight the need for a deeper understanding of YACS’ specific nutritional needs and more effective strategies to engage this cohort. Future research should prioritise larger, more representative samples, incorporate objective nutritional and clinical measures, and explicitly address psychosocial and age-related barriers to healthy eating in young adulthood. Full article

Portfolio optimization remains a challenging problem due to the dynamic, nonlinear, and uncertain characteristics of financial markets. Traditional portfolio construction approaches, including mean variance optimization and conventional Black–Litterman models, often suffer from inaccurate estimation of expected returns and unstable allocation caused by parameter uncertainty. These limitations become more significant under structural breaks, regime transitions, volatility clustering, and extreme market events. This study proposes a Transformer-GAN enhanced Black–Litterman framework (TG-BL) that integrates temporal representation learning, uncertainty-aware scenario generation, and Bayesian portfolio optimization. The proposed framework consists of three complementary components. First, a Transformer-based encoder is employed to extract long- range temporal dependencies and latent market representations from historical financial sequences. Second, a conditional Generative Adversarial Network (GAN) is introduced to generate diverse future return scenarios conditioned on Transformer- derived market representations, enabling probabilistic modeling of future uncertainty rather than deterministic prediction. Third, the generated return distributions are incorporated into the Black–Litterman framework through dynamically calibrated views and confidence estimation. Unlike conventional approaches that directly replace equilibrium returns with machine-generated predictions, the proposed method preserves the Bayesian structure of Black–Litterman by adjusting the influence of model- generated views according to predictive uncertainty. This mechanism allows AI- based forecasts to complement rather than dominate market equilibrium information. Extensive experiments are conducted using historical financial data under multiple market conditions. The evaluation framework includes portfolio performance comparison, GAN-generated scenario validation, robustness analysis under volatility and liquidity stress, and component- wise ablation experiments. The results demonstrate that the proposed TG-BL framework improves risk-adjusted portfolio performance while maintaining robustness against market uncertainty. The findings indicate that the integration of temporal feature extraction, uncertainty modeling, and Bayesian portfolio allocation provides an effective decision-support framework for quantitative investment management. Full article

As regulatory requirements increasingly shape automated lending decisions, fairness remains a critical challenge in high-stakes domains, particularly credit scoring. Although artificial intelligence models can achieve strong predictive performance, they may also reproduce biased outcomes that reduce financial inclusion or transfer harm to overlooked protected groups. Existing fairness interventions commonly operate at a single stage of the decision-making pipeline, despite bias often propagating across representational and decision layers. This study proposes iCert-Fair, a two-layer framework for technical fairness assessment and harm recovery in credit scoring. The first layer adopts a fairness-through-explainability paradigm, using SHAP-based explanations to identify direct and proxy dependence on protected attributes and guide structural dataset repair, while the second layer applies targeted threshold-policy adjustments to recover residual harm while preserving decision utility. Experiments on the German and Taiwanese credit datasets show that fairness gains are model- and dataset-specific and may be collective, concentrated, transferred, or recovered unevenly across protected attributes. The direct comparison with representative pre-processing, in-processing, and post-processing methods revealed that baseline methods targeting one protected attribute at a time frequently transferred residual harm to other monitored attributes. In contrast, the fairness-focused recommendations generated by iCert-Fair achieved larger collective fairness improvements across all considered protected attributes while avoiding residual harm. These gains were obtained while preserving predictive utility on the German dataset and with utility degradation remaining below 5% across the evaluated performance metrics on the Taiwanese dataset, alongside consistently lower false-negative risk. The empirical findings support the use of complementary structural and policy-level interventions and demonstrate the importance of jointly evaluating aggregate disparity, worst-case attribute-level harm, cross-attribute transfer, and predictive utility. Full article

Nigerian secondary schools face significant challenges integrating fourth industrial revolution (4IR) technologies. Little is known about how contextual factors influence teachers’ familiarity with these technologies and their perceptions of student engagement, learning outcomes, and instructional efficiency. Using a convergent parallel mixed-methods design, this study examined those factors in public and private secondary schools across one Local Government Area within each of four Nigerian states (Imo, Rivers, Osun, and Kogi). The quantitative component surveyed 189 teachers (92 public, 97 private). The qualitative component interviewed nine teachers and conducted two focus groups with 12 teachers (six in each group). Quantitative findings showed that teachers reported moderate familiarity with most 4IR technologies (grand mean = 3.11) and positive perceptions of their influence on student engagement (grand mean = 3.64) and learning outcomes (grand mean = 3.80). Private school teachers reported significantly higher familiarity with six of seven technologies and more positive perceptions of engagement (p < 0.001) and learning outcomes (p < 0.001) than public school teachers, except for WhatsApp (p = 0.332). Qualitative thematic analysis revealed four themes explaining these differences: (1) structural and infrastructural inequities (electricity, internet, devices) shape technology engagement; (2) teachers sustain integration through personal financial sacrifice and informal peer support; (3) professional development is perceived as contextually disconnected; and (4) technological engagement reflects interactions between generational confidence, pedagogical identity, and school culture. The joint display of quantitative and qualitative findings shows convergence on the conclusion that infrastructure, not teacher attitude, is the main barrier to 4IR adoption. Policymakers should prioritise investment in electricity, internet, and devices over blaming teachers. Full article

Artificial intelligence (AI) is reshaping the financial sector by redefining how customers access, interpret, and experience banking services. However, financial inclusion is still frequently analyzed in terms of access, use, or availability of digital channels, with less attention paid to the relational quality of interactions between customers and institutions. This study introduces Relational Inclusion in Banking (RIB) as a specific dimension of financial inclusion in AI-mediated contexts, referring to the degree to which customers continue to feel understood, treated fairly, heard, and connected to their institution. Drawing on Social Exchange Theory and Sociotechnical Systems Theory, a PLS-SEM model was tested using data from 770 banking customers in Extremadura, Spain. The results show that institutional trust was the strongest antecedent of RIB, followed by relational mediation capacity, perceived fairness of AI, and AI transparency. In addition, the rural-urban context influenced digital literacy level, although it did not directly explain institutional trust. At the theoretical level, this study contributes to existing knowledge by shifting the analysis of financial inclusion from digital access toward the relational position customers retain within automated banking services. It also identifies human mediation as a sociotechnical mechanism that connects AI-assisted outcomes with customers’ specific circumstances. At the practical level, the managerial implications direct financial institutions toward clear AI governance, better-designed AI-assisted interactions, and effective procedures for explanation and human review. The policy implications highlight the need to strengthen the traceability of automated decisions, ensure effective human review mechanisms, and monitor potential inequalities associated with territory or digital vulnerability. They also point to the value of maintaining hybrid support mechanisms that enable customers with greater digital difficulties to understand and challenge AI-assisted decisions. Full article

Credit risk assessment forms a cornerstone of banking risk management and the stability of the wider financial system. Over the past decade, the rapid development of machine learning (ML) techniques has substantially enhanced traditional credit risk assessment methodologies. ML has now emerged as a core technological pillar for the banking sector, strengthening risk identification capabilities, optimising credit decision-making, and advancing financial inclusion. Conventional credit scoring models, dominated by logistic regression (LR) and scorecard approaches, offer inherent strengths in interpretability and regulatory compliance. However, constrained by their linear assumptions, these methods struggle to capture complex non-linear relationships within credit data and deliver insufficient predictive accuracy for the “credit-invisible” population lacking formal credit histories. This paper presents a systematic literature review (SLR) of ML applications in credit risk assessment (CRA), covering publications from January 2016 to May 2026. A total of 894 papers were retrieved from five digital libraries, and following a rigorous multi-stage screening process, 129 studies were selected for final inclusion. Our analysis reveals that tree-based ensemble models and deep learning (DL) architectures predominate in contemporary research in this field. Meanwhile, post hoc explanation methods and machine learning operations (MLOps) are gaining significant traction as solutions to address fairness, transparency, and system maintenance challenges in real-world production environments. We synthesise prevailing methodologies into a unified end-to-end credit risk modelling framework spanning data preprocessing, feature engineering, model training, evaluation, and operational deployment. Through a critical assessment of the advantages, limitations, and inherent trade-offs of existing approaches, this SLR not only identifies current research gaps and future directions for the academic community, but also provides practical guidance for the banking sector to build compliant, fair, and efficient intelligent risk assessment systems. Full article

The management of oligometastatic disease has undergone a significant paradigm shift over the past two decades. Once considered uniformly incurable, selected patients with metastatic disease can now achieve prolonged progression-free survival, durable disease control, and, in carefully selected cases, long-term remission or cure through metastasis-directed therapies. Advances in stereotactic ablative radiotherapy (SABR), surgery, systemic therapies, and the emerging concept of Curative Oligometastatic Radiotherapy (CORT) have challenged the traditional distinction between curative and palliative treatment. Concurrent developments in imaging, including PET/CT, prostate-specific membrane antigen (PSMA) PET, whole-body MRI, and MR-guided adaptive radiotherapy (MR-linac), together with evolving biomarker research, are improving disease characterisation, refining patient selection, and treatment personalisation. As survival improves, an increasing number of patients are living with durably controlled metastatic cancer and experience long-term physical, psychological, cognitive, functional, and financial consequences of treatment. Despite these challenges, evidence-based survivorship pathways for patients with oligometastatic disease remain poorly defined. Supportive oncology is becoming an essential component of modern radiation oncology rather than an adjunct to cancer treatment. This emerging discipline focuses on optimising symptom control, minimising toxicity, and delivering structured survivorship care. Rather than being limited to end-of-life care, supportive oncology is embedded throughout the patient journey; from diagnosis and treatment selection to prehabilitation, rehabilitation, patient-reported outcome (PRO) monitoring, surveillance for late effects, multidisciplinary follow-up, and long-term survivorship. This review discusses how advances in precision radiotherapy, molecular imaging, biomarkers, and emerging treatment technologies are reshaping the management of oligometastatic disease while simultaneously creating a growing population of long-term survivors with increasingly complex supportive care needs. It highlights the expanding role of supportive oncology in the care of patients with oligometastatic disease, encompassing multidisciplinary symptom management and argues that improvements in disease control must now be matched by the development of evidence-based multidisciplinary survivorship pathways that integrate supportive oncology to optimise quality of life (QoL), functional independence, and patient-centred outcomes. Finally, this review highlights current evidence gaps and proposes future research priorities for developing evidence-based survivorship models for this rapidly expanding patient population. Full article

Climate change poses escalating risks to small-scale farmers’ food security across Sub-Saharan Africa (SSA), yet evidence on how adaptation strategies transform into food security outcomes remains fragmented and unevenly distributed. This study systematically reviewed 66 peer-reviewed studies retrieved from various databases including Sabinet African Journals, Web of Science and PubAg. The peer-reviewed studies were obtained through following PRISMA guidelines which involve searching criteria (using key words) and screening criteria (applying filters). The aim of this study was to synthesise climate-related challenges, adaptation strategies and policy gaps affecting small-scale farming systems in SSA. Empirical findings indicate that prolonged droughts, rainfall variability and temperature stress dominate climate-related challenges. Small-scale farmers’ adaptation responses cluster around technological and on-farm practices, integrated climate-smart agriculture, livelihood diversification, crop- and livestock-based strategies and knowledge-based interventions. As a result most strategies positively influence the availability, stability and access pillars of food security. However, adaptation effectiveness is strongly constrained by systemic policy gaps, notably weak extension and institutional support, limited access to finance and inputs, affordability and scalability barriers to CSA technologies. These constraints disproportionately affect resource-poor farmers, reinforcing inequalities and limiting the scalability of successful local adaptation. Therefore, strengthening institutional capacity, aligning policies with farmer realities, integrating local knowledge and gender considerations and expanding financial access emerge as central priorities for transforming adaptation efforts into sustained and equitable food security outcomes across SSA. Full article

Infrastructure project reports contain rich narrative evidence on delay causes, yet transforming such unstructured text into reliable causal knowledge remains challenging because reports mix confirmed events with hypothetical, conditional, or localized statements. This study proposes an eight-stage computational pipeline that converts infrastructure project evaluation reports into a Bayesian-network model for exploratory structure learning and probabilistic dependency modeling. The central methodological contribution is a modality-aware extraction layer that distinguishes confirmed, project-wide delay evidence from conditional, hypothetical, or component-level statements before causal analysis. The pipeline was evaluated on 55 road infrastructure project reports financed by the Asian Development Bank, the African Development Bank, and JICA, from which delay events across 15 cause categories were extracted and stratified by epistemic modality and scope. Ablation analysis shows that the principal dependency structure recovered by the Bayesian network is not recoverable without modality-aware filtering, indicating that evidence-quality stratification materially shapes downstream causal-structure exploration. Among the recovered dependencies, a financial-to-project-management pathway was the most consistent signal: its undirected skeleton edge was the only relationship recovered by all four causal-discovery algorithms tested (with the orientation determined only by the score-based search), its association was nominally positive—though weak and not uniformly discernible—across nine extraction models spanning three commercial vendors and open-weight families, and it is consistent with prior delay-factor literature. Its model-based scenario contrast ($\Delta P=+0.638$, 95% CI $[0.470,0.764]$) is reported as hypothesis-generating rather than as a validated policy effect: under structure-learning uncertainty, the interval extends to zero, and the effect magnitude and the specific learned edge depend on the extraction model and the small effective sample. These findings suggest that incorporating modality awareness into narrative-evidence extraction improves the reliability of exploratory causal-structure analysis from infrastructure project reports. Full article

Background/Objectives: End-of-life (EOL) preparedness remains critically understudied among Hispanic patients with advanced breast cancer and their patient–caregiver dyads, despite evidence that preparedness significantly influences quality of life, care decisions, and caregiver well-being. This study aimed to explore how Hispanic patient–caregiver dyads conceptualize and experience EOL preparedness. Methods: A qualitative descriptive design was employed, guided by the Dyadic Cancer Outcomes Framework, which highlights patient and caregiver characteristics, relationship processes, individual and relational outcomes, the cancer care trajectory, and the broader social context as interrelated influences on dyadic experience. Semi-structured individual interviews were conducted in Spanish with 11 metastatic patient–caregiver dyads (n = 22 participants) recruited through Ponce Health Sciences University and the Ponce Research Institute in Puerto Rico. Data were analyzed using codebook thematic analysis in NVivo 15, with themes interpreted through the framework’s components. Results: Six interdependent themes of EOL preparedness were identified: psychological and emotional, spiritual, informational, practical, physical, and caregiver role preparedness. Spiritual preparedness, grounded in faith, prayer, and surrender to divine will, functioned as the foundational axis organizing all other themes. Preparedness was dynamic and turning-point-driven, challenged anew at each stage of disease progression. Financial vulnerability, caregiver invisibility within formal care systems, and insufficient anticipatory information were identified as primary barriers. Family support and faith communities were the most consistently cited facilitators. Conclusions: The findings yield the first grounded, dyadic conceptualization of EOL preparedness with Hispanic advanced breast cancer patient–caregiver dyads. We propose a formal definition positioning preparedness as a dynamic, multidimensional, relationally embedded, and spiritually anchored process that is fundamentally interdependent between patient and caregiver. These results directly inform the development of a culturally tailored, dyadic EOL preparedness intervention for this underserved population. Full article

Stock market price prediction remains a persistent challenge owing to the non-stationarity, high noise content, and intricate spatiotemporal dependencies that characterize financial time series. Existing approaches typically excel at either local pattern extraction or long-range dependency modeling, yet seldom reconcile both within a unified framework. This paper introduces a CNN–Transformer dual-channel architecture equipped with a dynamic attention fusion module for stock price forecasting. The convolutional channel applies hierarchical dilated convolutions to distill fine-grained local patterns from multi-indicator sequences while suppressing high-frequency noise. Simultaneously, the Transformer channel employs multi-head self-attention to capture long-distance temporal correlations and regime-shift dynamics. A learnable gating mechanism then fuses the two feature streams by adaptively weighting local detail against global trend information according to market conditions. Experiments conducted on four real-world stock datasets spanning the S&P 500, CSI 300, NASDAQ Composite, and Hang Seng Index show that the proposed model reduces mean absolute error by 9.7–15.3% and root mean square error by 9.5–13.8% relative to competitive baselines including LSTM, CNN–LSTM, Informer, and PatchTST. Ablation studies further indicate that both channels and the fusion module contribute to prediction accuracy, and the architecture remains effective across markets with differing volatility profiles. Full article

Urban regeneration has become a primary strategy for addressing the spatial, social, environmental, and economic challenges associated with rapid urbanization. By using the PRISMA framework, this paper provides a systematic review of urban regeneration, focusing on its objectives, indicator systems, and estimation approaches. Journal articles published between 2008 and 2025 were retrieved from Web of Science, and 50 studies were retained after screening. The findings indicate that regeneration has increasingly highlighted long-term economic efficiency and environmental performance, along with growing attention to social equity, urban safety, and resilience. Life cycle costs are recognized as a core framework for capturing long-term financial and environmental performance. In terms of methodological evolution, a progression from conventional estimation techniques toward machine learning and hybrid approaches is observed. Several challenges remain, including the absence of standardized data frameworks and limited cross-regional transferability of existing models. To advance the field, future research could prioritize the establishment of common data standards, the adoption of explainable machine learning for improved model transparency, and the validation of models across diverse geographical and project settings. Full article

Under the backdrop of the global economic transition towards high-quality development, green innovation has emerged as a pivotal driver in the pursuit of Sustainable Development Goals. However, enterprises often face financing constraints and high-risk challenges when engaging in green technology innovation. As a novel financial model that integrates financial resources with industrial operations, supply chain finance (SCF) has emerged as a potential solution. This study aims to investigate whether SCF effectively promotes corporate green innovation and to analyze its underlying mechanisms. Drawing on data from China’s A-share listed firms, this study measures corporate green innovation using patent data, which are identified from the IncoPat database based on the WIPO Green Inventory’s IPC classifications. Concurrently, firms’ deployment and depth of engagement in supply chain finance are assessed via textual analysis. The empirical findings demonstrate that SCF exerts a significant positive impact on fostering corporate green innovation. Mechanism analysis further uncovers that SCF enhances green innovation capabilities primarily through two pathways: mitigating financial constraints and accelerating the process of digital transformation. Results from the heterogeneity tests demonstrate that such an incentivizing impact is predominantly pronounced among enterprises characterized by superior ESG performance. Furthermore, our empirical investigation reveals a distinct inverted U-shaped pattern characterizing the nexus between SCF and corporate green innovation, which implies that an excessive dependence on SCF could ultimately impede innovative activities. We provide firm-level empirical evidence that deepens the understanding of SCF’s role in fostering green innovation, revealing that SCF is a key instrument for financial resources to integrate into the green transition. Policymakers and managers should utilize and develop SCF within a reasonable scale and focus on digital upgrading to achieve sustainable economic growth. Full article