Search Results (304)

The impact of chain-leading (hub) firms’ forward-looking information disclosure on supply chain collaborative innovation remains underexplored. Based on panel data of Chinese A-share manufacturing listed firms from 2014 to 2024, this study adopts large-scale textual analysis to construct a time-varying indicator measuring hub firms’ forward-looking disclosure, and systematically examines its innovation spillover effects and internal mechanisms. The results show that hub firms’ forward-looking disclosure is positively associated with a significant increase in the R&D investment intensity of supply-chain node firms. This spillover effect is negatively moderated by node firms’ information absorptive capacity, reflecting a typical information substitution effect. Heterogeneity tests further reveal that the spillover effect is more pronounced among node firms with larger scale, higher supply-chain network centrality, and stronger supply-chain relationship specificity (proxied by higher customer concentration). In addition, such innovation spillovers are positively associated with improved corporate financial performance, and this profit-conversion effect is more pronounced among high-leverage firms, which is consistent with an implicit endorsement mechanism that helps alleviate financing constraints. Combining empirical evidence with industrial governance practice, this paper expands the theoretical boundary of supply chain collaborative innovation and provides actionable recommendations for optimizing information disclosure rules and formulating differentiated industrial innovation policies. Full article

Background: The integration of Artificial Intelligence (AI) and algorithms into healthcare is transformative, particularly in diagnosing rare diseases (RDs), enhancing the accuracy and speed of condition identification. Objectives: This systematic literature review investigates perceptions and attitudes toward the use of AI in Electronic Health Records (EHRs) for screening patients at risk of RD, aiming to understand patients’ and healthcare providers’ expectations and concerns. Methods: Following PRISMA guidelines, a systematic search was performed in December 2023. A search strategy developed by the research team in collaboration with an expert librarian, using the PICO framework, was applied. Searches were conducted in PubMed, Scopus, and Web of Science. The search strategy covered four main concepts: diagnostic techniques, medical records, AI, and attitudes toward these technologies. Results: The initial search retrieved 3348 articles after duplicate removal. However, no studies met the inclusion criteria. As a result, no eligible studies were identified, preventing risk-of-bias assessment or data synthesis. Discussion: The absence of relevant studies highlights the need for further research focusing on patient and healthcare provider attitudes toward AI-integrated EHRs, especially in RD and their early detection. Conclusions: The lack of studies on stakeholder attitudes toward AI in EHRs for RD screening represents an important research gap. Addressing this gap will improve the understanding and development of AI applications in healthcare, ensuring they meet user needs and ethical standards. Full article

This paper investigates the dynamic coopetition and capacity-sharing strategies between an Integrated Manufacturer and a Developer under R&D uncertainty, focusing on the governance of strategically scarce capacity. By constructing a two-stage game model, we analyze how government intervention and risk-hedging mechanisms influence the allocation of idle strategically scarce capacity in innovation-driven industries. The findings reveal a two-sided paradoxical behavioral pattern: in the low-probability R&D interval, rather than relying on safe contract manufacturing, the Integrated Manufacturer counter-intuitively reduces collaborative duration to aggressively gamble on its immature product. Conversely, in the high-probability R&D interval, where conventional wisdom predicts an aggressive pivot to self-production, the manufacturer paradoxically extends or maintains the contract manufacturing duration, driven by the partner’s full cost-sharing incentive mechanism. Furthermore, To maximize the total supply output of strategically scarce resources during collaboration, we uncover a non-linear ‘counterproductive subsidy trap’ and propose a binary ‘critical mass’ policy rule: governments should either withhold subsidies entirely or commit sufficient funding to bypass the supply deficit zone. This framework provides a theoretical foundation for managing scarcity in capital-intensive sectors such as biopharmaceuticals and semiconductors. Full article

Introduction: Hypertension (HTN) management is essential to improving cardiovascular and cerebrovascular morbidity and mortality in the older adult, yet optimal blood pressure (BP) targets remain uncertain. We aim to compare the incidence of major adverse cardiovascular events (MACE) among older adults presenting to the emergency department (ED) with systolic blood pressure (SBP) 140–159 mmHg and diastolic blood pressure (DBP) 90–99 mmHg versus those with SBP ≥ 160 mmHg and DBP ≥ 100 mmHg. Methods: Retrospective analysis was conducted using the Global Collaborative Network including adults aged ≥ 65 years presenting to the ED with essential HTN without end-stage renal disease. Cohort 1 (SBP ≥ 160 mmHg and DBP ≥ 100 mmHg) included older adults with more severely elevated BP, while Cohort 2 (SBP 140–159 mmHg and DBP 90–99 mmHg) included those with moderately elevated BP. The primary outcome was all-cause mortality over 5 years. Secondary outcomes included AKI, CHF, AMI, ischemic stroke, and hemorrhagic stroke. Propensity score matching was used to balance baseline characteristics. Results: After propensity score matching, 191,829 patients remained in each cohort. The mean age was 72.6 (±5.9) years; 51.8% were female, 10.5% had diabetes mellitus, and 5.65% were obese. Cohort 2 (SBP 140–159 mmHg) had lower risks across all primary and secondary outcomes compared with Cohort 1 (SBP ≥ 160 mmHg). The largest risk difference (RD) was observed for AKI (RD = 1.28%; 95% CI, 1.09–1.46%; p < 0.0001), followed by ischemic stroke (RD = 1.25%; 95% CI, 1.11–1.39%; p < 0.0001), CHF (RD = 1.11%; 95% CI, 0.93–1.29%; p < 0.0001), AMI (RD = 0.70%; 95% CI, 0.58–0.83%; p < 0.0001), all-cause mortality (RD = 0.33%; 95% CI, 0.15–0.51%; p < 0.0001), and hemorrhagic stroke (RD = 0.29%; 95% CI, 0.22–0.36%; p < 0.0001). All outcomes demonstrated consistently higher absolute risk in Cohort 1 compared with Cohort 2. Conclusion: Cohort 2 (SBP 140–159 mmHg) was associated with a statistically significant but small decrease in cardiovascular, cerebrovascular, and mortality outcomes, with all absolute RDs uniformly small (<1%) and minimal survival benefit over time. These small gains must be weighed against overtreatment risks, such as medication costs, hypotension, and falls, underscoring the need for individualized, risk-based HTN management in older adults. Full article

This study provides a systematic examination of global research trends and developments in the field of family and marriage over a twenty-five-year period (2000–2025). Employing a hybrid review design, the research integrates bibliometric analysis with PRISMA guidelines to ensure methodological rigor and transparency. Data were retrieved from the Web of Science, where an initial pool of 97,171 records was refined to 2974 eligible publications through a structured screening and inclusion process. The reduction to 2974 publications was the result of structure bibliometrics using CiteSpace, which employs algorithmic thresholds to identify the most structurally significant publications within a large corpus. Utilizing CiteSpace (version 6.4.R1), this analysis maps the intellectual structure and evolution of the field. By synthesizing co-citation, co-authorship, institutional, and keyword co-occurrence data, this study identifies critical collaboration networks, influential contributors, and dominant thematic domains. The findings reveal prominent research clusters, including premarital cohabitation, partner effects, family structure transitions, marital discord, systemic family functioning, and marriage education. Key contributors identified include influential scholars such as Catherine Walker O’Neal, Birditt, Kira S, Higginbotham Brian J, Beach Steven R. H., and Matthew D. Johnson. Leading institutions are the University System of Ohio, the University of California System, the Pennsylvania Commonwealth System of Higher Education (PCSHE), Pennsylvania State University, and Pennsylvania State University–University Park. At the country level, the United States, Canada, England, Australia, the Netherlands, and Belgium emerge as the most significant contributors. The findings offer a comprehensive synthesis of authorship trends, institutional influence, and shifting research trajectories within the field of family and marriage studies. Full article

Objective: Glutamic acid decarboxylase 65 (GAD65) antibody-associated epilepsy often presents as chronic focal epilepsy, usually with temporal lobe predominance, marked drug resistance, and inconsistent response to first-line immunotherapy. We assembled a large, harmonized, and literature-derived patient-level cohort to examine whether immunotherapy timing and regimen composition were associated with seizure outcome and to identify clinically meaningful prognostic signals. Methods: We performed a literature-derived patient-level analysis of 375 unique published cases linked to 132 contributory source publications from an audited full-text register of 166 reviewed studies. Descriptive analyses used the whole cohort. Treatment-response analyses assessed seizure outcome at the first evaluable post-immunotherapy assessment and at the last follow-up. Good seizure outcome was defined as seizure freedom and/or ≥50% seizure reduction. The primary timing comparison contrasted early treatment, defined as immunotherapy within 6 months of symptom onset, with late treatment, defined as immunotherapy after more than 12 months; four cases treated in the intermediate >6 to ≤12 month window were retained for descriptive timing summaries but excluded from the primary comparison. Statistical testing used the Fisher exact, Chi-square, Mann–Whitney U, and prespecified clustered logistic sensitivity analyses where appropriate. Results: The pooled phenotype was predominantly female, usually temporal-lobe-based, and frequently drug-resistant, with common autoimmune comorbidity and heterogeneous MRI abnormalities. Among timing-evaluable treated cases, earlier immunotherapy showed a class-specific, exploratory signal rather than a uniform regimen-independent effect. In rituximab/CD20-directed regimens, early treatment was associated with a higher rate of good seizure outcome than late treatment at both the first post-immunotherapy assessment and last follow-up (93.8% vs. 50.0%; risk difference [RD]: 43.8 percentage points; 95% CI: 7.7 to 72.7). A similar pattern was observed in the broader escalation group (94.4% vs. 55.6%; RD: 38.9 percentage points; 95% CI: 6.3 to 68.1). By contrast, steroid-containing regimens showed no clear early-versus-late advantage (84.6% vs. 88.2%; RD: −3.6 percentage points; 95% CI: −18.4 to 20.1). Shorter epilepsy duration before immunotherapy and absence of established drug resistance were the most clinically meaningful favorable baseline features. Significance: In GAD65 antibody-associated epilepsy, the therapeutic window may be most relevant for escalation strategies rather than for steroid-containing first-line regimens. However, these class-specific findings are exploratory and hypothesis-generating. They derive from non-randomized, literature-derived data and may reflect treatment intensity, center practice, publication era, and confounding by indication rather than isolated regimen superiority. Prospective collaborative registries with standardized longitudinal seizure outcome measures are needed to validate these observations. Full article

This paper examines how technology blockade affects leader and follower firms’ research and development (R&D) incentives and their cooperation decisions under asymmetric knowledge spillovers, while also exploring the role of government subsidies in mitigating market failures and restoring cooperation incentives. Motivated by the increasing restrictions on knowledge diffusion in high-technology industries, we develop a two-stage game in which firms first choose R&D investment and then compete in quantities under both non-cooperative and cooperative regimes. Our analysis shows that the impact of technology blockade on firms’ R&D investment and profit distribution depends on R&D efficiency and the presence of asymmetric knowledge spillovers. Specifically, under non-cooperative behavior, the interaction between asymmetric spillovers and R&D efficiency generates nonlinear effects on both R&D efforts and profit allocation. Under cooperative regimes, although firms can internalize spillovers, technology blockade reduces coordination benefits and leads to asymmetric profits, resulting in the absence of a self-enforcing cooperation region. Furthermore, our results indicate that government subsidies can partially or fully restore cooperation incentives, thereby increasing R&D investment and enhancing social welfare in most cases. These findings highlight a substitution effect between policy intervention and external technological constraints, emphasizing the importance of targeted subsidies in mitigating the adverse effects of technology blockade on innovation and collaboration. Full article

Rapid and often chaotic urbanisation in post-Soviet cities such as Almaty challenges the quality, availability, and accessibility of public spaces for residents, given the cities’ historical development. Meanwhile, global research is focused on the concepts of Third Places, coworking spaces in the Western context, and urban experience in cities with transitional economies, but the heritage of centrally planned urban development lacks spatial explicit analysis. The purpose of the current study is to analyse the evolution, current situation, and distribution of collaborative spaces (public spaces that combine work and connectedness) in Almaty. The methodology includes four phases of qualitative analysis: (1) a historical–typological analysis of architectural functions since the beginning of the 20th century until the 2025, (2) spatial mapping analysis of the existing typologies such as libraries, museums, coworking spaces, research and development (R&D) institutions and universities, and community centres, (3) longitudinal statistical analysis, and (4) historical graphic analysis. Analysis is conducted through the lens of advanced levels of human needs that concern self-education and self-development. This approach helped to propose a new definition of collaborative space. The results also show examples of sustainable urban structure with collaborative spaces in Almaty’s old centre (“Zolotoi Kvadrat”—Golden Square) and a critical deficit of new multifunctional spaces for work and socialisation in recently developed districts. The study reveals that Almaty’s evolution occurred through incremental infill development over the old grid, without the integrated development of the public realm and existing structural connections. As a result, the research explores the connection between collaborative spaces and their indirect influence on the general well-being in Almaty. Full article

This experimental study investigated the impact of embedding social and emotional learning (SEL) in science instruction on the academic and social–emotional outcomes of Arabic-speaking sixth graders, including those with reading difficulties (RD). Children from two schools in northern Israel (n = 101) were randomly assigned to either an intervention group, which received SEL-enriched science lessons featuring collaborative, reflective activities over 30 sessions, or a control group receiving traditional instruction. Pre- and post-tests assessed SEL competencies, motivation towards science, and academic achievements in science and mathematics. Results showed significantly greater gains in SEL skills, and in science motivation and science achievement in the intervention group compared to controls, whereas mathematics outcomes remained unchanged. Typically developing students and those with RD benefited similarly. Integration of SEL into science curricula thus enhances cognitive and social–emotional learning dimensions, particularly in linguistically and socio-economically marginalised populations. Implications for inclusive pedagogy and future research directions are discussed. Full article

Background/Objectives: Spontaneous bacterial peritonitis (SBP) is a serious complication of decompensated cirrhosis and is associated with acute kidney injury (AKI), organ failure, and death. Intravenous albumin is recommended in SBP because it reduces renal impairment and mortality, particularly in patients at higher risk of circulatory dysfunction and hepatorenal complications. However, the prognostic impact of early albumin administration on clinical outcomes in hospitalized SBP patients remains incompletely characterized in real-world practice. This study aimed to assess the association between early albumin administration and clinical outcomes in patients hospitalized with SBP compared to those without early albumin. Methods: A retrospective cohort study was conducted using the TriNetX US Collaborative Research Network, including adults hospitalized with SBP through February 2026. Patients were divided into those receiving early albumin administration (n = 1248) and those without early albumin (n = 4932) within 24 h of index SBP diagnosis. Propensity score matching (1:1) balanced cohorts (n = 1230 each) for demographics, comorbidities, liver disease severity surrogates, medications, and laboratory values. Relative risks (RR), risk differences (RD), and hazard ratios (HR) were calculated using propensity-matched and Cox proportional hazard models. Results: Early albumin administration was associated with significantly lower all-cause mortality (RR 0.620; 95% CI: 0.441–0.871; p = 0.005 at 5 days; RR 0.770; 95% CI: 0.651–0.910; p = 0.002 at 90 days). Secondary outcomes showed reduced risks for acute kidney injury (RR 0.654; 95% CI: 0.553–0.774; p < 0.001 at 5 days; RR 0.798; 95% CI: 0.706–0.903; p < 0.001 at 90 days), hepatorenal syndrome–AKI (RR 0.598; 95% CI: 0.445–0.804; p < 0.001 at 5 days; RR 0.756; 95% CI: 0.613–0.932; p = 0.009 at 90 days), vasopressor requirement (RR 0.633; 95% CI: 0.489–0.820; p < 0.001 at 5 days; RR 0.712; 95% CI: 0.572–0.887; p = 0.002 at 30 days), and renal replacement therapy (RR 0.533; 95% CI: 0.324–0.878; p = 0.011 at 5 days; RR 0.642; 95% CI: 0.442–0.932; p = 0.019 at 30 days). Cox models confirmed statistically significant risk reductions for all primary and secondary outcomes, including ICU admission (HR 0.82; 95% CI: 0.73–0.92; p = 0.001) and 30-day readmission (HR 0.84; 95% CI: 0.73–0.97; p = 0.015). Associations were strongest in the early period and attenuated over time. Conclusions: Early albumin administration was associated with reduced risks of mortality, AKI, HRS-AKI, and hemodynamic instability in patients hospitalized with SBP, with attenuation over time. These findings support timely implementation of guideline-concordant albumin therapy, although residual confounding cannot be excluded. Full article

Vegetation phenology refers to the cyclical growth patterns of vegetation in nature, which are influenced by climatic conditions, human activities, and genetic factors. It plays an irreplaceable role in regulating carbon cycling and energy flow within natural ecosystems. However, the combination of a cloudy and rainy climate with a landscape characterized by the interplay of land and water and fragmented patches has long posed challenges for remote sensing phenological monitoring data, including a scarcity of valid observations, frequent temporal gaps, and spectral distortion in mixed pixels. These issues make it difficult to reliably support the needs of wetland phenological inversion and mapping. To address this issue, this study uses vegetation inversion in the Poyang Lake wetlands as a case study and reconstructs high-spatiotemporal-resolution time-series kNDVI data based on multi-source remote sensing data. Methodologically, we propose an improved and enhanced spatiotemporal adaptive reflectance fusion model, IESTARFM. This model enhances the homogeneity of similar pixel selection through adaptive matching windows and land cover constraints. Additionally, it explicitly incorporates cloud probability and time-lag factors into the weighting structure to systematically downweight unreliable observations, and further employs quadratic term corrections to account for the nonlinear growth response of kNDVI. Using the reconstructed dataset, key phenological information is extracted by combining third-order harmonic analysis with a dynamic thresholding method, thereby enhancing the robust characterization of seasonal trajectories under conditions of missing data and noise. Accuracy evaluation results show that the 10m/8d high-frequency kNDVI dataset reconstructed by IESTARFM achieves at least a 12.61% improvement in fusion accuracy compared to classical methods such as ESTARFM, STARFM, and FSDAF, with a maximum reduction in RMSE of 0.026, and effectively restores details in areas with thin cloud cover. The reconstructed kNDVI series achieved a coefficient of determination $R^2$ = 0.875 and RMSE = 0.066 relative to Sentinel-2 observations, indicating that the reconstructed series closely reproduces the reference imagery in both amplitude and spatial structure. The phenological parameters derived from kNDVI exhibit an RMSE of 4.81 days compared to field observations, demonstrating that the reconstructed time series reliably captures the timing of key phenological events. It should be noted that the proposed approach is designed for post-event time-series reconstruction and is not intended for real-time forecasting. In summary, this study collaboratively enhanced the reliability of high-resolution index time-series reconstruction and phenological identification in cloudy and rainy wetlands through three key aspects: cloud noise suppression, heterogeneous boundary preservation, and nonlinear growth characterization. It provides a generalizable technical foundation for dynamic monitoring of wetland vegetation, ecological restoration assessment, and refined management in regions with frequent cloud and rainfall. Full article

This study presents a bibliometric and systematic review of 480 articles meeting the following inclusion criteria: English-language articles, reviews, or proceeding papers focusing on building topics with full text available, retrieved from the Web of Science Core Collection on 9 Jannary 2026 to map the intellectual landscape of smart-sustainable building (SSB) research. Employing the PRISMA framework combined with scientometric mapping (VOSviewer), thematic classification, and qualitative synthesis (no risk of bias assessment was performed as this was a bibliometric review), the analysis reveals exponential publication growth since 2022, identifying three dominant thematic clusters: digital enabling technologies (41.0%), energy systems (30.8%), and advanced building envelopes and materials (28.3%). Keyword analysis identifies “smart buildings,” “green buildings,” and “energy efficiency” as central conceptual anchors, while temporal trends indicate increasing attention to artificial intelligence, digital twins, and blockchain. Notably, 51.4% of articles address two or more themes simultaneously, confirming the field’s interdisciplinary character. Critical analysis reveals persistent fragmentation: sustainable building rating tools (e.g., BREEAM, LEED) and smart building evaluation methods (e.g., Smart Readiness Indicator). Seven challenges, including assessment fragmentation, high costs, and cybersecurity vulnerabilities, are identified as barriers to SSB adoption. Limitations include reliance on a single database (Web of Science) and subjective thematic classification. This review provides a roadmap for future research emphasizing integrated assessment frameworks and interdisciplinary collaboration. Registration: Not pre-registered. Funding: National Key R&D Program of China (2025YFF0521003). Full article

To promote the green and low-carbon transition and achieve sustainable development in the transportation sector, virtual calibration technology was employed for the efficient and precise control of emissions from heavy-duty diesel engines and aftertreatment systems. A data-driven, semi-empirical and semi-physical simulation modeling method was proposed. By constructing core modules based on physical mechanisms and refining empirical parameters using experimental data, the method improves computational efficiency while maintaining the prediction accuracy of key parameters. Additionally, a collaborative architecture combining physical actuators and virtual sensor signals was introduced, laying the foundation for the validity of virtual calibration. By innovatively introducing a closed-loop system with real actuators and virtual sensors, the dynamic response characteristics of the control system are faithfully reproduced, providing a reliable environment for validating the results of virtual calibration. Under steady-state conditions, the results demonstrated an average relative error of 1.7% for brake-specific fuel consumption (BSFC) and 6.1% for NOx emissions. An open-loop system for the virtual calibration testing platform was constructed for steady-state calibration. Using the main injection timing and common rail pressure as independent variables, a D-optimal design was utilized to generate 43 sets of experimental points, from which a polynomial regression model was established (R² ≥ 98%). Under the constraints of NOx and pre-turbine temperature, fuel consumption in the low-load range is reduced by 0.5–3 g/kW·h, aftertreatment NOx emissions are reduced by 0.5–3 g/kW·h, and exhaust temperature is increased by 10 °C. This study establishes a complete development workflow consisting of “operating condition design-virtual optimization-bench validation,” significantly enhancing calibration efficiency and engineering applicability. This method shortens the calibration cycle and reduces the number of physical bench tests, providing the industry with a comprehensive calibration methodology tailored to engine operating conditions that is both reproducible and scalable. Full article

Surface rock movement can lead to geological or environmental problems such as surface subsidence, ground fissure development, and deformation of engineering structures, and its evolution process exhibits significant spatiotemporal heterogeneity. Therefore, conducting high-precision, spatiotemporally continuous monitoring of surface deformation is of great significance for revealing subsidence mechanisms, assessing potential risks, and guiding disaster reduction decisions. GNSS and InSAR have become mainstream methods for monitoring surface deformation, but they still have limitations in terms of spatial sparsity, 3D deformation inversion capability, and data gaps in areas of strong deformation. To address these issues, this paper takes the Jinchuan copper-nickel mine’s No. 2 mining area as the research object and comprehensively utilizes multi-source monitoring data from GNSS and InSAR to construct a joint inversion model of the surface 3D deformation field based on posterior variance component estimation, achieving adaptive optimization of weight allocation and collaborative solution of 3D deformation. To address the issue of InSAR decorrelation in areas of strong deformation, which leads to missing deformation information, a fitting and estimation approach was applied to supplement six decorrelated points that spatially coincide with GNSS stations. These points are located in key deformation areas, and their reconstruction effectively improves the completeness and reliability of the deformation field in critical regions. Based on this, an automated solution process for the fusion model is implemented using MATLAB R2022b, and the joint inversion yields spatiotemporally continuous 3D deformation fields in the northward, eastward, and vertical directions. The results show that compared with traditional monitoring methods, the proposed fusion model exhibits higher inversion accuracy and stability under different InSAR technology conditions, effectively suppressing the impact of single data source errors on the overall solution results. Among them, SBAS-InSAR shows slightly higher accuracy in the vertical direction, while PS-InSAR achieves higher accuracy in the planar direction, as indicated by lower RMSE and MAE values. The research results improve the accuracy and reliability of surface deformation monitoring in mining areas, providing important technical support for safe mining and refined management. Full article

To address the problem in predicting target prices for ship equipment orders where small sample sizes, high feature dimensions, and strong business constraints lead traditional models to overfit and have insufficient generalization ability, a method combining GAN and NVP-D for small-sample data augmentation and price prediction is proposed. This method integrates the advantages of adversarial training in Generative Adversarial Networks (GAN) with the explicit density estimation and stable training characteristics of Normalizing Flow NVP-D. By using dual-weight collaborative optimization of the objective function, it alleviates gradient vanishing and mode collapse, generating high-quality virtual samples that closely follow the real data distribution. Redundant features are removed using Boruta-Lasso joint feature selection to reduce model complexity. CatBoost is employed as the prediction model to complete price estimation. Experiments were conducted on a ship equipment dataset with 33 original samples and 24 features, strictly following the standard procedure of augmentation only within the training set and 5-fold cross-validation. Compared with NVP, NVP-G, MAF, traditional GAN, and Mixup methods, the results show that the proposed integrated model achieves optimal performance when augmenting 400 samples, with an RMSE of 0.0675, MAE of 0.0510, and R² of 0.9228. After feature selection, prediction accuracy further improves, with RMSE decreasing to 0.0615 and R² increasing to 0.9341. Limited by the scale of the original samples, the statistical robustness and cross-dataset generalization capability of this method still need validation with larger datasets. However, under the current small-sample constraints, it can effectively alleviate modeling bottlenecks and provide high-precision support for equipment procurement argumentation, budget preparation, and cost control in stages. Full article