Search Results (74,074)

Consumer complaint narratives are an underused source of business and regulatory evidence. This study presents a proof-of-concept analytical framework for multi-dimensional complaint analysis and uncertainty-aware evidence synthesis, with potential use in human-in-the-loop decision support, and applies it to 99,434 Consumer Financial Protection Bureau (CFPB) complaint records spanning from March 2015 to March 2026. The empirical workflow integrates BERTopic topic discovery, two pretrained sentiment checkpoints (ProsusAI/finbert and cardiffnlp/twitter-roberta-base-sentiment-latest), exploratory clustering of complaint records, and descriptive temporal summaries. A conceptual stock–flow and feedback representation links observed complaints, analytical alerts, response capacity, and unresolved issues; these relations are not causally estimated or simulated. BERTopic produced 89 non-outlier topics. A probability-based outlier-reassignment stage assigned 80.46% of records to these topics, while 19.54% remained unassigned and were retained as an uncertainty queue. A FinBERT-prediction-stratified benchmark of 600 complaint records was independently coded by two annotators (raw agreement = 87.5%; Cohen’s κ = 0.758). Against Annotator 1 as the prespecified reference, Cardiff RoBERTa aligned better with the labels on the prediction-stratified benchmark (κ = 0.395; sample accuracy = 67.3%) than FinBERT (κ = 0.195; sample accuracy = 46.3%), although neither model supports autonomous use. The four-group K-means solution is reported as exploratory because the highest observed Silhouette coefficient was only 0.106 and favored K = 2. Temporal patterns coincided with selected external events, but no causal effect is claimed. The contribution is therefore a transparent, uncertainty-aware analytical architecture whose potential organizational uses require human review and operational validation. Full article

The diffusion of information-and-communication technologies (ICT) and, increasingly, artificial intelligence (AI) across higher education—the transition from ICT-supported teaching toward intelligent, data-driven, and sustainable educational applications—has produced a large but fragmented literature, and no systematic account yet maps the thematic structure of this AI–ICT intersection quantitatively. This study addresses that gap with a systematic literature review coupled to a reproducible topic-modeling pipeline. Following the PRISMA 2020 protocol, we searched six bibliographic databases and screened the results down to a corpus of 126 peer-reviewed journal articles published between 2021 and 2026. The corpus was analyzed with a Latent Dirichlet Allocation model whose number of topics and hyper-parameters were selected by a many-objective NSGA-III optimization that jointly balances topic coherence, separation, distributional equity, generalization, and stability, complemented by an overlap-aware soft assignment that quantifies cross-cutting studies. The analysis recovers three comparably sized themes—institutional adoption and digital transformation; generative AI, instructional design, and critical thinking; and AI for teaching quality, performance, and intelligent systems—each accounting for a comparable share of the corpus. Interpreted through an ICT–AGI framework used as an analytical lens, scholarly attention spreads across the framework’s administrative, generative-AI, and learning-facing dimensions rather than concentrating on any one, and about a quarter of the studies bridge more than one theme; the strongest coupling links institutional adoption to teaching quality and the weakest links generative AI to teaching quality, a difference that a null-model test shows is not statistically robust, so we report it as suggestive rather than as a firmly established gap. The study offers an evidence-based thematic map of the AI–ICT intersection in higher education and a transferable, many-objective methodology for machine-assisted evidence synthesis. Full article

Ground-based microwave radiometers (MWRs) continuously retrieve atmospheric temperature and vapor density profiles, but systematic biases relative to radiosonde observations limit their quantitative application in nowcasting and data assimilation. In this study, we matched temperature and vapor density profiles from six ground-based MWRs with L-band radiosonde observations at Wuhan to evaluate data quality for 2025 and analyze the sources of station-to-station bias differences. We established linear regression, random forest (RF), and artificial neural network (ANN) models to correct biases by height group and sky condition, respectively, and compared the corrected data with the original data. The results show that, compared with radiosonde data, the MWR temperature was generally cold-biased under clear (−0.895 °C), cloudy (−0.729 °C), and rainy (−0.188 °C) skies, whereas the vapor density was moist-biased under all three conditions (+0.417, +0.209, and +0.47 g/m³). The per-station biases differed markedly and could not be explained by spatial separation alone, indicating station-specific environmental and representativeness differences rather than instrument differences. All three correction methods significantly reduced the deviations; taking the RMSE together with the residual bias as the joint criterion, the ANN was selected as the best method, reducing the test set temperature RMSE from 2.802 °C to 2.284 °C (−18.5%) and the vapor density RMSE from 2.764 g/m³ to 1.746 g/m³ (−36.8%), followed by random forest (−17% and −35.9%) and linear regression (−6.8% and −6.8%). After removing the small constant residual bias, the ANN is both the most accurate and effectively unbiased. The corrected profiles brought the thermodynamic instability parameters to a usable level for severe-convection nowcasting and provided an early-warning signal in the pre-onset period of gale and short-duration heavy rain, which also provide a reference for the quantitative application of microwave radiometer data. Full article

At present, government-led research and other scientific institutions continue to face challenges in meaningfully weaving Indigenous Knowledge in research, thereby reinforcing the inequalities and exclusion faced by Indigenous Peoples. We found that, among other reasons, these challenges stem in part from a lack of tools and resources to guide researchers in engaging Indigenous partners ethically and respectfully throughout the research design, practice, and reporting phases, rather than to a lack of awareness or motivation. In response to this need, we developed an Inclusive Science Self-assessment Tool (ISSAT), an online questionnaire-based toolkit designed to support weaving Indigenous Knowledge Systems (IKS) in scientific research. ISSAT was developed for IKS through a three-year process that combined evidence synthesis with dynamic Indigenous rights-holder and stakeholder engagement. The evidence synthesis comprised a systematic review of ten years of relevant literature and a case analysis of how 215 federal scientists weave IKS in their research. Through the dynamic stakeholder approach, we engaged a cross-section of participants and experts across government, academia, and civil society whose input informed the development and validation of the tool. ISSAT users complete the questionnaire and receive tailored feedback with practical strategies for fostering inclusivity in scientific research, prompting self-reflection through an inclusive and relational research ethics lens. At the institutional level, the tool enables institutions to identify areas of need and prioritize limited resources to drive the institutional change needed for an inclusive, equitable, and diverse scientific enterprise. Full article

Accurate multi-step satellite cloud-image prediction remains challenging because cloud fields exhibit displacement, deformation, and local appearance changes over time. This study proposes M-SimVP, a multi-channel prediction framework built on SimVP. A Motion-Aware Temporal Prediction (MATP) module models temporal changes and learns alignment fields from historical features, while a Multi-Scale Motion Compensation (MSMC) module transfers the learned alignment information to multi-scale features and image space. A generation-refinement pathway further complements deformation-based compensation. Experiments on FY-4B/AGRI CH1-CH3 sequences use six input frames to predict four future frames at 15-, 30-, 45-, and 60 min lead times. M-SimVP achieves an MSE of 0.00421, SSIM of 0.7604, PSNR of 23.76 dB, and Edge-MSE of 0.0180, reducing MSE and Edge-MSE by 12.29% and 16.67%, respectively, compared with SimVP. The model also outperforms recurrent, optical-flow/advection, and satellite-specific baselines. Ablation, channel-wise, and lead-time analyses confirm the contributions of the proposed components and show that the performance advantage is maintained across the 1 h forecast horizon, while complex cloud evolution remains challenging at longer lead times. Full article

Optimizing semen collection is critical for maximizing genetic advancement and reproductive efficiency in livestock. While exogenous hormones can enhance ejaculate quantity and quality, short windows of biological activity often limit their efficacy. This study was conducted in northern-arid Mexico (25° N, 103 W) to evaluate the comparative effects of oxytocin (OXYT) and carbetocin (CARB) administration on semen quality in French-Alpine bucks. Males (n = 12) were allocated into three groups and received intravenous treatments 10 min prior to artificial vagina collection: (1) OXYT (10 UI oxytocin buck⁻¹), (2) CARB (0.2 mg 100 kg⁻¹ LW⁻¹), and (3) CONT (1 mL saline buck⁻¹). The same bucks received all three treatments separately. CARB outperformed both groups (p ≤ 0.05), improving all evaluated parameters (p ≤ 0.05): ejaculated volume (i.e., increased by 103%), sperm concentration, mass and progressive motility, total sperm count, viability, normal morphology, plasma membrane functionality, acrosome integrity, and chromatin dispersion. Crucially, CARB doubled (p ≤ 0.05) the available projected or theoretical number of doses for intravaginal artificial insemination (POTNDIAI), followed by the OXYT and CONT groups, with 11.14 POTNDIAI each. Therefore, CARB emerges as a highly effective, accessible pharmacological strategy for semen collection protocols, substantially enhancing sperm yield and structural quality in breeding bucks. Ultimately, by bridging advanced reproductive physiology with practical smallholder farming, this methodological optimization provides a viable strategy to enhance reproductive efficiency and long-term sustainability in small ruminant production systems within arid environments. Full article

Background/Objectives: Adnexal torsion causes ovarian ischemia/reperfusion injury and may impair ovarian reserve. This study evaluated the protective effects of Fetuin-A in a rat torsion/detorsion model. Methods: Twenty-eight female Wistar rats were assigned to Sham, T/D, T/D+Fetuin-A treatment, and Fetuin-A pretreatment+T/D groups. Bilateral ovarian torsion was induced for 3 h followed by 2 h of detorsion/reperfusion, and Fetuin-A was administered intraperitoneally at 100 mg/kg. Serum cytokines and AMH, qRT-PCR markers, histopathological injury, and immunohistochemical expression of Fetuin-A, HIF-1α, GPx, and Bcl-2 were evaluated. Results: T/D induced a systemic inflammatory response, reduced AMH, increased HMGB1/TLR4-associated and cell-stress markers, and caused marked ovarian histopathological injury. Fetuin-A attenuated inflammatory and molecular stress responses, while pretreatment produced the most consistent protection, preserving AMH, reducing histopathological injury and HIF-1α immunoreactivity, and restoring GPx, Bcl-2, and Fetuin-A immunoreactivity. Exploratory STRING analysis identified functional enrichment related to oxidative stress, autophagy, apoptosis, and HIF-1 signaling, providing hypothesis-generating context without establishing causal pathway activity. Conclusions: These findings indicate that Fetuin-A administration, particularly before torsion, was associated with reduced inflammatory and tissue-stress responses during early ovarian ischemia/reperfusion injury. Further studies are needed to determine optimal dosing, clinically feasible timing, and long-term reproductive outcomes. Full article

Objective: Rheumatoid arthritis (RA) is a chronic autoimmune disease with the abnormal accumulation of pro-inflammatory macrophages within the synovial joints. Total flavonoids of Rhizoma Drynariae (TFRD), the major bioactive constituents of the traditional Chinese medical herb Drynariae rhizome, have shown potent activities in treating RA, but their effect and underlying mechanism in modulating macrophage behavior remain unknown. This study aimed to evaluate the therapeutic potential of TFRD in regulating macrophage polarization during RA, and to further elucidate the underlying mechanism. Methods: The therapeutic effects of TFRD were assessed by utilizing a collagen-induced arthritis (CIA) mice model in vivo. A macrophage polarization model was employed to investigate the effect of TFRD on the macrophage pro-inflammatory polarization and glycolytic activity in vitro. Additionally, the role of hypoxia-inducible factor-1α (HIF-1α) in the TFRD-mediated regulation of macrophage glycolysis was examined via plasma transfection. Results: Administration of TFRD significantly reduced the arthritis severity scores and ameliorated joint swelling and bone destruction in CIA mice. The therapeutic efficacy of high-dose TFRD (TFRD-H) was broadly similar to that of MTX in CIA mice. Further, TFRD decreased the proportion of pro-inflammatory macrophages and the expression of HIF-1α within the joints of CIA mice. The in vitro experiments indicated that TFRD inhibited the lipopolysaccharide-induced macrophage pro-inflammatory polarization. Mechanistically, TFRD attenuated glycolytic metabolism and downregulated HIF-1α expression during pro-inflammatory polarization, while HIF-1α overexpression partly reversed these effects. Furthermore, 16S rRNA sequencing analysis revealed the mitigated dysbiosis of the gut microbiota and increased relative abundance of Parabacteroides in CIA mice. Conclusions: These findings indicate that TFRD alleviated RA, which was partly dependent on downregulating pro-inflammatory macrophages and mitigating HIF-1α-associated glycolysis. This study could provide a scientific basis for further research and the clinical use of TFRD. Full article

This prospective single-center cohort study evaluated changes in serum levels of novel adipokines and myokines, including isthmin-1 and irisin, as well as the advanced glycation end product (AGE)–soluble receptor for AGE (sRAGE) axis and monomeric C-reactive protein (mCRP) in patients with obesity before and 5 months after laparoscopic sleeve gastrectomy (LSG). The study included 38 patients with obesity (26 females and 12 males) with a median age of 42 years who underwent LSG. Laboratory parameters were assessed using commercially available immunoenzymatic assays. We observed a significant decrease in isthmin-1 concentration at the postoperative time point (Me = 2.10 ng/mL vs. Me = 1.93 ng/mL, p = 0.009). Moreover, a significant increase in irisin levels (Me = 7.33 µg/mL vs. Me = 10.64 µg/mL, p = 0.002) and AGEs (Me = 589.50 ng/mL vs. Me = 698.60 ng/mL, p = 0.015) was observed after LSG. A significant increase in AGE levels was also noted after normalization to total protein (Me = 8.34 ng/mg protein vs. Me = 10.37 ng/mg protein, p = 0.027). sRAGE levels were similar at both time points. Five months after LSG, mCRP levels were significantly lower (Me = 1.99 ng/mL vs. Me = 1.55 ng/mL, p = 0.020). Exploratory correlation analyses identified potential associations between AGEs and isthmin-1 (R = 0.37) and between sRAGE and irisin (R = −0.31) preoperatively. Five months after LSG, potential associations were observed between AGEs and irisin (R = 0.38) and between sRAGE and mCRP (R = −0.37). Our findings indicate a decrease in circulating isthmin-1 and mCRP, together with an increase in irisin and AGE concentrations following LSG, suggesting changes in selected metabolic and tissue-related biomarkers after surgery. Further studies are needed to determine the mechanisms underlying these changes and to clarify their potential clinical significance. Full article

Morocco’s wheat sector faces a widening imbalance between rising domestic demand and climate-sensitive domestic supply. This study assesses wheat production and food security in Morocco over 1990–2023 through regional and national analyses and scenario-based projections to 2035 and 2045. Regionally, wheat production declined between 2010 and 2020, except in Casablanca-Settat, where irrigation supported a 161% production increase despite drought. Nationally, production increased significantly (+75,880 t yr⁻¹, p = 0.02), despite 2012–2023 being the warmest and driest decade on record. Harvested wheat area and fertilizer inputs (NPK) were retained as predictors of national production, together yielding an adjusted R² of 0.52, with harvested area being the strongest statistical predictor (adjusted R² = 0.53 when considered alone) compared with NPK (adjusted R² = 0.12). Scenario-based projections show the wheat food security deficit increasing from 4.54 Mt in 2023 to 7.11 Mt by 2045. Theoretical calculations indicate that closing this deficit would require allocating an additional 19–27% of available cropland to wheat under irrigated conditions, compared with 45–63% under rainfed conditions. These findings underscore the potential role of irrigation and the need for region-specific agricultural strategies under a warming and drying climate. While the present analysis relies primarily on national-scale statistics and linear relationships, future work using finer-resolution data and complementary modeling approaches, together with explicit consideration of water-resource availability, could further refine the assessment of irrigation expansion and its feasibility. Full article

Local short-duration heavy-rainfall warning requires early assessment of hazardous rainfall within a specified neighborhood. Geostationary meteorological satellites provide frequent, spatially continuous observations of cloud systems and storm evolution, but only indirectly reflect surface rainfall. Rain gauges measure surface accumulation directly, but only at irregular locations. Deep learning fusion methods commonly convert gauge observations to gridded representations before combining them with satellite observations. This treatment merges gauge measurements and local gauge support into a single representation, limiting how local surface conditions inform warning decisions and increasing the likelihood that local heavy-rainfall risk is underestimated or overestimated. We present PLUME (Precipitation Nowcasting with Late-Stage Residual Updates from Rain-Gauge Measurements for Short-Duration Heavy-Rainfall Events), a staged fusion model with gridded and point-origin gauge pathways. PLUME first combines a gauge-derived gridded rainfall background with FY-4A AGRI observations to form a base rainfall representation. The point-origin pathway uses local gauge features and observation support to update this representation through signed residuals, which the station decoder then maps to neighborhood-event probabilities. This staged design preserves the complementary roles of continuous spatial context, local rainfall measurements, and gauge availability. We evaluated PLUME for 0–3 h local short-duration heavy-rainfall warning over central and eastern China using an independent test set from May to September 2023. Across Barnes analysis, inverse distance weighting, and kriging, PLUME consistently improved the critical success index (CSI) over matched satellite–grid baselines. Relative CSI gains were 2.6–3.3% under complete gauge input and 3.6–5.9% under simulated gauge missingness. PLUME also reduced the associated CSI loss by 45.8–67.4%. Analysis of the variant with frozen batch normalization (BN) statistics showed that, under all three backgrounds and both input conditions, the largest positive probability updates were concentrated among baseline misses and converted some to hits, whereas the largest negative probability updates were concentrated among baseline false alarms and converted some to correct negatives. The dominant conditional CSI contribution shifted from false-alarm suppression at 0–1 h to missed-event recovery at 1–3 h. By using gridded and point-origin gauge information at different fusion stages, PLUME improves local heavy-rainfall warning through lead-dependent bidirectional refinement. Full article

User stories are central to agile requirements engineering, but in practice they often exhibit ambiguity, missing information, and weak or absent acceptance criteria, thereby increasing rework and reducing verifiability. This paper presents LOONA, a mathematically formulated NLP/LLM decision-support pipeline for Spanish user story quality assurance. We model the task as supervised binary classification followed by conditional text generation: a fine-tuned Spanish BERT classifier estimates whether a story requires improvement, and a generative model is triggered only when the estimated class exceeds a decision threshold. The study uses an anonymized reconstruction of 100 Spanish-language user stories from an administrative software backlog, labeled by three agile requirements experts through majority vote. The reconstructed annotation table yields Fleiss’ $\kappa =0.76$, indicating substantial agreement. We report stratified 5-fold classification results, lexical baselines, rule-based baselines, automatic rewrite-similarity metrics, and paired expert review of held-out outputs. LOONA achieved Accuracy $=0.91$, Precision $=0.90$, Recall $=0.92$, F1-score $=0.91$, and MCC $=0.82$, outperforming TF-IDF logistic regression, TF-IDF support vector machines, and a rule-based baseline. In paired expert review ($n=30$), 70% of generated rewrites improved over the original story, 20% were unchanged, and 10% degraded; generated acceptance criteria were judged correct in 86.7% of cases and free of hallucinated information in 93.3%. The contribution is framed as a formal and empirical proof of concept for model-level LLM-assisted requirements quality assurance, with explicit limitations regarding dataset size, confidentiality, and human oversight. Full article

Background: Post-tuberculosis lung disease (PTLD) is associated with persistent functional impairment and reduced quality of life. This study evaluated the association between the addition of low-level laser therapy (LLLT) and deep oscillation (DO) to structured pulmonary rehabilitation and changes in health-related quality-of-life and functional outcomes in patients with clinically stable PTLD. Methods: In this prospective, controlled, non-randomized study, 65 patients completed an eight-week program. The study group (SG; n= 33) received structured kinesiotherapy plus LLLT and DO, while the control group (CG; n= 32) received the same kinesiotherapy alone. Between-group effects were assessed using baseline-adjusted ANCOVA with HC3 robust standard errors and Benjamini–Hochberg correction. Results: Greater improvements were observed in the combined-intervention group across all prespecified outcomes (all q < 0.001). Adjusted between-group differences were 6.46 points (95% CI 5.73–7.19) for EUROHIS-QOL-8, 8.86 (7.39–10.33) for WHOQOL-BREF physical health, 14.69 m (10.57–18.81) for 6MWD, −0.62 s (−0.94 to −0.30) for TUG, −1.04 (−1.29 to −0.79) for mMRC, and −2.21 (−2.63 to −1.79) for Borg CR10. Conclusions: Adjunctive LLLT and DO were associated with greater improvements in quality-of-life and functional outcomes than kinesiotherapy alone. Given the non-randomized design and the combined administration of the two adjunctive modalities, these findings should be interpreted as associative and require confirmation in randomized, sham-controlled trials. Full article

This study analyzes Kazakhstan’s tourism competitiveness through a descriptive and comparative design, drawing on the WEF Travel and Tourism Development Index (TTDI), the World Bank Logistics Performance Index (LPI), and national and international strategic documents. Kazakhstan possesses relative advantages in transport infrastructure and natural resources compared to Russia, China, Kyrgyzstan, and Uzbekistan. However, it exhibits significant weaknesses in tourism service infrastructure, human resource development, and promotion. To address these weaknesses, the study proposes the “KAZTUR-4 Strategic Framework,” which recombines selected dimensions of Porter’s Diamond Model (1990), Dwyer and Kim’s Integrated Model (2003), and Ritchie and Crouch’s Destination Competitiveness Model (2003) through an explicit mapping procedure that links Kazakhstan’s empirically identified weaknesses to theoretical constructs. The framework comprises four components: transport infrastructure; human resources and educational development; tourism product diversification and infrastructure; and promotion and marketing. Demand conditions, destination management, sustainability, safety, and local welfare are treated as cross-cutting dimensions. Within the TTCI series, Kazakhstan moved from 93rd (2011) to 79th (2019); the 2021 TTDI rank of 66th belongs to a methodologically revised index and cannot be read as a continuation of this trajectory. Tourism’s GDP contribution stood at only 0.8% in 2019, and the average length of stay was 2.6 nights—the lowest in the region. Uzbekistan and Kyrgyzstan moved more decisively on visa liberalization, as reflected in their TTDI 2021 International Openness scores (78.4 and 82.1 vs. Kazakhstan’s 42.5). Uzbekistan also outperformed Kazakhstan on Prioritization and Promotion (52.1 vs. 48.2). Full article

Artificial intelligence (AI) is generating unprecedented technological opportunities while raising profound ethical, anthropological, and social challenges. This article examines the role of the Catholic Church in shaping AI ethics within the tension between technological progress and a personalist understanding of the human person. The study employs a hermeneutical approach and critical analysis of the scholarly literature and key documents of the Church’s Magisterium, including Gaudium et Spes, Laudato si’, Laudate Deum, and the Rome Call for AI Ethics. It also compares Catholic ethical reflection with selected secular frameworks, namely the OECD Principles on Artificial Intelligence, the UNESCO Recommendation on the Ethics of Artificial Intelligence, and the European Union AI Act. The findings indicate that Catholic reflection on AI is grounded in an integral personalist anthropology that emphasises human dignity, the common good, moral responsibility, and the relational nature of the person as fundamental criteria for evaluating technology. While secular regulatory models primarily focus on safety, transparency, and accountability, the Catholic perspective contributes a deeper anthropological and theological dimension rooted in the concept of imago Dei. The article proposes a model of ‘personalist ethics of co-creation’ that integrates contemporary AI ethics standards with a Christian vision of human responsibility. Such an approach may complement existing legal and technical regulations and promote a more human-centred development and use of AI. Full article