K-means Research Articles

Abstract 2348: Unveiling dichotomies between liver cancer and adjacent normal tissues in Wilson’s disease model

Abstract Objective: The Long-Evans Cinnamon (LEC) rat is a model for Wilson’s disease (WD) and is associated with chronic inflammation and hepatocellular carcinoma. This study performed a gene expression microarray using total RNA extracted from LEC liver tissue at 84 weeks late chronic hepatitis, tumor-adjacent normal (84 weeks N), and 84 weeks tumor (84 weeks T). Understanding the difference between a tumor and adjacent normal within the same animal can be very intriguing as both are in the same environment, and one becomes and sustains a tumor status, but the other still maintains its normalcy. Experiment: Total RNA was extracted from liver tissue of three LEC rats at 84 weeks N and 84 weeks T and hybridized to the Affymetrix GeneChip Rat Genome 230 2.0 Array (Affymetrix Inc, Santa Clara). GeneGO Pathways software was used to analyze the differentially expressed genes using a cutoff of P-value ≤0.05 and FDR <0.25% and a fold change +/- 2. A volcano plot was generated to visualize significant gene change expression between 84 weeks N vs 84 weeks T. Agglomerative Hierarchical and K-means clustering were used to group similarly expressed genes into clusters and build a hierarchy of clusters. Each cluster obtained from K-means clustering analysis within the up or down-regulated groups were checked for pathway analysis to understand the interplay between biological pathways by QIAGEN Ingenuity® Pathway Analysis (QIAGEN IPA). In addition − log(p) > 1.3, a z-score cut-off͕ ≥ 2 for activation and a z-score cutoff ≤ -2 for inhibition was used to call a pathway activation, and inhibition is significant. Result: The seven most significant pathways that were activated are the S100 Family Signaling Pathway, Extracellular Matrix Organization, Response to Elevated platelet cytosolic Ca2+, HMGB1 Signaling, Sertoli Cell-Sertoli Cell Junction Signaling, Cachexia Sunaina Pathway, and Nicotine Degradation II. Except for Nicotine Degradation II pathway, all other pathways are shown activated in non-alcoholic fatty liver disease (NAFLD) and hepatocellular carcinoma (HCC) patients previously but shown in the WD model for the first itme. The Nicotine Degradation II pathway is shown to be activated in the NAFLD patient before. Conclusion: To our knowledge, this is the first report where gene expression analysis was done between tumor-adjacent normal and tumor within the same animal in the WD model. Our results identified multiple pathways in the model that could lead us to new diagnostics and therapeutic targets when validated. Citation Format: Ritam Adhikari, Chiranjeev Dash, Rabindra Roy. Unveiling dichotomies between liver cancer and adjacent normal tissues in Wilson’s disease model [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2348.

SmartICST: a smart information and cyber security training approach for older adults

Abstract An increasingly aging population, widespread internet connectivity, and an increasing number of cybersecurity incidents bring many challenges for the future of information security. Older adults, in particular, are a highly vulnerable demographic, often targeted by cyber attackers. This paper aims to develop an approach that enables the identification of different groups of users of information systems according to their requirements for information and cyber security training (ICST). To achieve this, we surveyed 784 older adults residing in the UK recruited through the Prolific platform. First, S-HAIS-Q (Shortened Human Aspects of Information Security Questionnaire) was used to assess their information security performance. Through hierarchical clustering and k-means clustering, we identified five distinct groups. Additionally, predictive clustering trees (PCTs) were employed to simultaneously perform clustering and classification, providing deeper insights into the characteristics of these groups. This approach provides a segmentation framework highlighting the potential for customizing ICST to address the diverse needs of different groups of older adults. The results of our study indicate that it is possible to tailor and optimize ICST even among older adults who are more proficient in computer use than the general older adult population.

Abstract 2051: Baseline and adaptive activated kinome profiling identifies PTK6 as a mediator of resistance to vemurafenib and dabrafenib but not encorafenib in BRAFV600-mutant metastatic melanoma

Abstract To investigate the landscape of the activated kinome in MM, we applied a proteomic approach that interrogates the activation status of more than half of the entire kinome (∼300), termed multiplexed kinase inhibitor bead affinity chromatography coupled with mass spectrometry (MIB/MS) on 45 MM patient (pt) tumors (BRAFV600, n=25; NRASQ61, n=5, NF1 loss-of-function, n=4; Triple wild-type n=11) in conjunction with whole exome sequencing and RNA sequencing (RNAseq). We performed exploratory analyses (Wilcoxon rank sum test for each kinase) on the untreated tumors to identify differentially activated kinases in BRAFV600-mutant MM vs. all others. While no activated kinases were differentially expressed in BRAFV600-mutant MM after adjustment for multiple testing, AKT1, PDGFRA, and MAP3K1 kinases exhibited higher MIB binding (i.e., more activated) in BRAFV600E-mutant pts (unadjusted p-value < 0.05). Spearman’s rank correlation tests between the activated kinase and its gene expression by RNAseq and adjusted for multiple testing did not identify a significant correlation for any kinase, suggesting a complex mechanism of kinase activation other than sole gene expression. K-means clustering analysis to classify melanomas into 4 subtypes using a 6-gene discriminant expression signature showed that BRAFV600-mutant tumors were more frequently classified as melanocytic (9/24, MLANA/MITF/SOX10high, AXLlow). In contrast, triple wild-type tumors were more frequently classified as transitory (5/10, MITF/SOX10/ETV4high, AXLlow). We then performed similar exploratory analyses on the baseline (untreated) tumors from the subset of the 20 BRAFV600E-mutant pts who received D+T as part of a prospective clinical trial (NCT01726738; median follow-up 30.5 months, range 2.0-116.1 months) to identify differentially activated kinases that would predict shorter (<12 months) vs. longer progression-free survival. Again, while no kinases survived multiple comparison testing, the bromodomain protein TRIM28, known to regulate melanoma plasticity, exhibited a more consistent loss of MIB binding in BRAFV600E-mutant pts who progressed shorter than in pts who progressed longer than 12 months on D+T (unadjusted p-value < 0.05). 7 baseline-progression tumor pairs were evaluated for MIB/MS, and 6 were for RNAseq. Although diverse resistance mechanisms were identified, as previously described, we also identified melanoma subtype switch in 4/6 pt tumors and a consistent loss of MIB binding of the non-receptor tyrosine kinase PTK6 in response to D+T, an off-target of D and vemurafenib but not encorafenib. Our in vitro findings revealed that off-target PTK6 inhibition by D could activate beta-catenin signaling and contribute to the development of resistance. Citation Format: Steve P. Angus, Naim U. Rashid, Andrew J. Walther, David L. Corcoran, Steven D. Rhodes, Frances A. Collichio, C. Paige Jones, Mikaela J. Bauer, Joel S. Parker, Noah Sciaky, 1Alisha R. Coffey, Samantha M. Bevill, Brian T. Golitz, v Timothy J. Stulhmiller, Nancy E. Thomas, David W. Ollila, Norman E. Sharpless, Carrie B. Lee, Gary L. Johnson, Stergios J. Moschos. Baseline and adaptive activated kinome profiling identifies PTK6 as a mediator of resistance to vemurafenib and dabrafenib but not encorafenib in BRAFV600-mutant metastatic melanoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2051.

High Concordance of E-Nose-Derived Breathprints in a Healthy Population: A Cross-Sectional Observational Study

Exhaled breath analysis using electronic noses (e-noses) is a promising non-invasive diagnostic tool. However, a lack of standardized protocols limits clinical implementation. This study evaluates the consistency of breathprints in healthy subjects using the Cyranose 320 e-nose to support standardization efforts. Breath samples from 139 healthy non-smoking subjects (age range 18–65 years) were collected using a standardized protocol. Participants exhaled into a Tedlar bag for immediate analysis with the Cyranose 320. Principal Component Analysis (PCA) was used to reduce data dimensionality, and K-means clustering grouped subjects based on breathprints. PCA identified four principal components explaining 97.15% of variance. K-means clustering revealed two clusters: 1 outlier and 138 subjects with highly similar breathprints. The median distance from the cluster center was 0.21 (IQR: 0.18–0.24), indicating low variability. Box plots confirmed breathprint consistency across subjects. The high consistency of breathprints in healthy subjects supports the feasibility of standardizing e-nose protocols. These findings highlight the potential of e-noses for clinical diagnostics, warranting further research in diverse populations and disease cohorts.

Mapping Left-Behind Regions: A Classification of Regional Disparities in Türkiye

This study examines regional disparities in Türkiye through the concept of "left-behind places," which has gained prominence in discussions on economic marginalization and political discontent. Despite the extensive literature on left-behind regions in the U.S., U.K., and Europe, Türkiye remains understudied. Using a Left-Behind Index (LBI) and K-means clustering analysis, this research identifies and classifies Türkiye’s NUTS2 regions based on long-term economic and demographic trends. The findings reveal persistent regional inequalities, characterized by an East-West divide and the economic underperformance of certain regions. Unlike Europe, where deindustrialization plays a crucial role, Türkiye’s left-behind regions are primarily defined by economic stagnation, population decline, and outmigration. The results underscore the need for targeted regional policies to address disparities. This study contributes to the broader discourse on left-behind places by providing empirical insights into Türkiye’s unique socio-economic landscape.

Years of potential life lost due to cardiovascular diseases in the Zabaykalsky Krai

Aim. To study the years of potential life lost (YPLL) of the population of the Zabaykalsky Krai due to cardiovascular disease (CVD) in the context of municipal structures for the development and justification of managerial measures.Material and methods. To calculate the standardized YPLL from CVD for the region as a whole and for municipalities separately, we used depersonalized data from form № S-51 "Information on the deceased by sex, age and cause of death", information on the average annual sex and age composition of the population from form 4RN of the Zabaykalsky Krai Statistics Service for the period of 2018-2022. Relatively homogeneous groups of municipalities by the level of standardized YPLL were identified using k-means cluster analysis. To test the significance of differences between clusters, variance analysis was performed.Results. When analyzing the YPLL, in the period of 2018-2022, the region's population annually lost from 30300 to 33989 person-years due to cardiovascular mortality. The total YPLL due to CVD increased by 0,2% over time. Throughout the analyzed period, the YPLL among men remained higher than among women, indicating excess mortality of the male working-age population. Cluster analysis based on YPLL coefficients standardized by age divided all region districts into 3 following clusters: with high, medium and low values . According to variance analysis, a significant difference was obtained between the studied clusters (p<0,05).Conclusion. The conducted study allowed us to identify the YPLL due to premature cardiovascular death in the Zabaykalsky Krai. Population cohorts and districts with high and increasing YPLL values were identified. The data obtained require management decisions in medical organizations aimed at improving preventive measures to promote a healthy lifestyle among the population.

Intelligent Clustering and Adaptive Energy Management in Wireless Sensor Networks with KDE-Based Deployment

Wireless sensor networks (WSNs) are widely used in IoT, environmental monitoring, and industrial systems, but ensuring energy efficiency, extended network lifetime, and reliable communication under real-world constraints remains challenging. This work proposes a novel clustering framework that integrates kernel density estimation (KDE)-based adaptive node deployment, silhouette-optimized K-means clustering, Bayesian cluster head (CH) selection with Gaussian noise-based energy uncertainty modeling, energy-efficient coverage control, and carrier sense multiple access with collision avoidance-based data transmission. Unlike conventional approaches that rely on fixed clustering and uniform deployments, our framework supports terrain-aware node placement and dynamic CH selection based on residual energy and distance under imperfect sensing conditions. Simulation results demonstrate significant improvements in performance, including over 35% extension in network lifetime and higher coverage retention under energy constraints, compared to baseline methods such as LEACH and K-LEACH. While detailed metrics vary per run due to adaptive parameters and stochastic node behavior, these outcomes validate the scalability, robustness, and practical relevance of the proposed method for real-world WSN deployments.

The Role of Personal Values in Sports Participation Among Young People with Disabilities: A Cluster and Profile Analysis

Sports participation among young people with disabilities offers significant physical, psychological, and social benefits, yet participation rates remain lower than among their non-disabled peers. This study, conducted in Hungary, explores how value orientations, health perceptions, and disability characteristics interact to shape sports engagement. Using a cluster analysis approach, we identified three distinct groups based on how young individuals with disabilities prioritize various life aspects. Data were collected through a survey of 771 participants aged 8–18, including individuals with diverse disability types. K-means clustering revealed three profiles: (1) a low sports activity group with moderate health consciousness, (2) a health-conscious group with broad life engagement but lower sports participation, and (3) a sports-oriented group that demonstrated strong interest in sports despite reporting lower self-rated health. The findings suggest that personal value orientation is a stronger predictor of sports participation than disability type or perceived health status. Furthermore, access to sports facilities alone does not guarantee participation, emphasizing the need for motivational and psychological interventions. These results highlight the importance of tailored, value-driven strategies in promoting physical activity among young people with disabilities, shifting focus from disability-specific adaptations to broader engagement-based approaches.

Habitat Radiomics and Deep Learning Features Based on CT for Predicting Lymphovascular Invasion in T1-stage Lung Adenocarcinoma: A Multicenter Study.

Habitat Radiomics and Deep Learning Features Based on CT for Predicting Lymphovascular Invasion in T1-stage Lung Adenocarcinoma: A Multicenter Study.

Clustering-guided prototypical contrastive learning for bearing fault diagnosis under variable working conditions

Abstract Methods based on deep learning for intelligent fault diagnosis have shown good results in general diagnostic tasks. Nevertheless, these methods largely depend on the sufficient labeled data, limiting their application in the actual scenarios where the availability of labeled data is limited. Moreover, the distribution of testing data is inconsistent with that of training data because bearings operate in various working conditions, leading to the performance degradation of these approaches. To tackle these two entangled problems, we propose a novel unsupervised domain adaptation network, which presents clustering-guided prototypical contrastive learning for cross-domain fault diagnosis. More specifically, k-means clustering is first used to aggregate similar source samples and target samples separately, acquiring the centroid of each cluster and the cluster index of each sample. Then, we propose in-domain and cross-domain contrastive learning strategies based on clustering results to achieve class alignment and domain alignment across source domain and target domain. By applying in-domain contrastive learning, we make the intra-class distance smaller while making the inter-class distance larger within each domain, effectively reducing the number of samples on the class boundaries. By applying cross-domain contrastive learning, class-to-class semantic similarity across two different domains is considered, which not only retains class discriminability in each domain but aligns these two domains at both the class level and the domain level. Detailed experiments on three bearing datasets reveal that our method outperforms in fault diagnosis across diverse working conditions, achieving average accuracy improvements of 2.10%, 7.44%, and 1.17% on the JNU, HUST, and Ottawa datasets, respectively.

A systematic review to identify target genes that modulate root system architecture in response to abiotic stress

The exposure of plant roots to soil-related stresses, including drought, high temperatures, salinization, and nutrient deficiency, is on the rise due to climate change caused by human activities. A systematic literature review was conducted, which revealed evidence for conserved genes that modulate root system architecture under specific stress conditions. A collection of Arabidopsis thaliana mutants displaying a root phenotype distinct from the wild type is available in The Arabidopsis Information Resource database. Gene expression data was gathered for specific genes in response to selected abiotic stress treatments. K-means clustering, and fold change analyses identified 118 genes that were upregulated and 185 genes that were downregulated. A dedicated phenotyping approach was used to ascertain that lack of nutrients induced the transition from a ‘steep, cheap, and deep’ root morphotype to a ‘topsoil foraging’ root morphotype in the Columbia- 0 reference genotype. The anticipated role of ISOPENTENYLTRANSFERASE 3, LIPOXYGENASE 1, and WEE1 KINASE HOMOLOG as negative regulators of root growth in response to multiple stress signals was assayed. Further research with the candidate genes identified in this study may reveal promising targets for crop improvement.

Assessment of water scarcity as a risk factor for cholera outbreaks.

Assessment of water scarcity as a risk factor for cholera outbreaks.

Quantitative Analysis of Cryptocurrency Susceptibility: A Mathematical Benchmarking Model

The rise of cryptocurrencies has brought attention to significant security challenges, particularly the 51% attack. This study focuses on developing benchmarks to evaluate varying levels of vulnerability among cryptocurrencies. A detailed review of the literature identifies a lack of approaches having statistical rigor, leading to the development of a comprehensive susceptibility test model. The proposed model is based on key parameters extracted from existing studies and validated with additional quantitative data for accuracy and reliability. Benchmarking thresholds are determined using k-means clustering, allowing for the classification of cryptocurrencies into distinct security profiles. The analysis identifies five clusters: resilient cryptocurrencies have susceptibility scores below 0.532, while scores exceeding 1.557 indicate high vulnerability. The remaining clusters represent intermediate levels of resilience and risk. These findings contribute to a better understanding of cryptocurrency security, supporting informed investment decisions and providing a basis for future research and policy development.

Study on air-rail intermodal ticketing optimization based on K-means clustering A case study of Qingdao Jiaodong International Airport

To improve the intermodal service at Qingdao Jiaodong Airport, addressing operational challenges such as fuzzy passenger demand layering and insufficient cross-modal coordination, and to solve the core issues of supply-demand mismatches and a single pricing mechanism in the air-rail intermodal ticketing system, this study proposes a personalized ticketing optimization strategy based on user profiling. First, through extensive survey data, the study analyzes the personal attributes and travel characteristics of the surveyed passengers. Then, using the K-means clustering algorithm, the study clusters passengers' multidimensional features and determines the optimal number of clusters through the elbow method and silhouette coefficient method. This leads to the establishment of differentiated user labels: economy-class passengers, business-class passengers, and leisure-class passengers. The market segmentation research on passenger groups shows that these three distinct groups perceive the bottlenecks of intermodal services differently, especially exhibiting significant layering features in the key dimensions of time sensitivity and price sensitivity. The results provide a comparative scheme for improving the air-rail intermodal ticketing service at Qingdao Jiaodong International Airport, offering differentiated service strategies for each passenger group. Through responsive demand and resource optimization, this study has significant practical implications for enhancing passenger experience and strengthening the market competitiveness of the service.

Clinical Clusters for Identification of Lower Adherence in Patients with Parkinson’s Disease: A Cross-Sectional Study

Abstract Background Non-adherence to pharmacotherapy in Parkinson’s disease (PD) is associated with worse clinical outcomes and poor quality of life (QoL). Early identification of non-adherent patients is crucial, as appropriate interventions can improve clinical conditions and QoL. Our study aimed to use cluster analysis to identify risk profiles of patients with lower adherence rates. Methods We included 124 cognitively intact patients with PD. Validated diagnostic instruments were used to measure adherence, QoL, non-motor symptoms (NMS), motor involvement, and complications. K-Means clustering was employed to create empirical subtypes based on these variables. Results Cluster analysis identified four distinct PD subtypes. Subtype 1 was characterized by worse motor state, frequent NMS, and poor QoL without complications. Subtype 2 had higher LEDD and complications but lower scores in other parameters. Subtype 3 showed low scores across all parameters, indicating a relatively good clinical condition. Subtype 4 showed higher scores in all observed parameters. Adherence levels significantly differed between subtypes (Cramer's V = 0.262, p = 0.009), with subtypes 1 and 4 showing lower adherence and subtype 3 showing higher adherence. Conclusion Patients with worsened motor state, NMS, and complications are more prone to lower adherence, which correlates with poorer QoL. Early identification and targeted interventions are essential to enhance adherence in these groups.

IMAGE PROCESSING FOR DETECTING MELANOMA SKIN CANCER USING AN OPTIMIZED ROTATION–INVARIANT COORDINATE CONVOLUTIONAL NEURAL NETWORK

Cancer is a dangerous disease caused by the unregulated proliferation of bodily cells. Cancer can grow in any part of the human body and may contain billions of cells. Skin cancer is a common type of cancer that originates on the top layer of the skin. Melanoma is one category of skin cancer, this is the most dangerous disease and leads to mortality. Generally, the machine learning approaches are mostly employed to diagnose skin cancer using protein sequences and other forms of imaging. But some disadvantages occur while using the machine learning systems, like requiring human-engineered features, difficulty, and being time-consuming. In this manuscript, Image Processing for detecting Melanoma Skin Cancer using an optimized Rotation–Invariant Coordinate Convolutional Neural Network (IP-MSCD-RICCNN) is proposed for skin cancer detection. The input images are gathered through the International Skin Imaging Collaboration (ISIC) dataset. Then the images are pre-processed using the Gillijn De Moor Filter (GDMF) to remove the noise. The pre-processed image is given to the Generalized Intuitionistic Fuzzy C-Means Clustering (GIFCMC) for segmenting the Region of Interest (ROI) from the skin image. The segmented image is supplied to the Newton Time-Extracting Wavelet Transform (NTEWT) to extract Gray Level Co-occurrence Matrix (GLCM) features such as Contrast, Correlation, Entropy, and Inverse. After feature extraction, the Rotation–Invariant Coordinate Convolutional Neural Network (RICCNN) is used to classify the skin images as malignant melanoma and benign melanoma. RICCNN generally does not adapt any optimization approaches to determine the optimal parameters for exact prediction. Therefore, the Mutated Coati Optimization Algorithm (MCOA) is employed to enhance the RICCNN for skin image classification. The proposed IP-MSCD-RICCNN method is implemented in Python. The proposed method achieves 26.36%, 20.69%, and 35.29% higher accuracy, 32.85%, 19.63%, and 25.47% higher precision, and 16.57%, 21.47%, and 30.56% greater sensitivity when compared to the existing models: skin cancer diagnosis from dermoscopy imageries utilizing deep learning with fuzzy k-means clustering (DI-SCD-DL-FKC), automatic detecting skin cancer by consolidating texture features with convolutional neural networks in dermoscopy images (DI-SCD-CNN-TF), and detection of skin cancer utilizing combined decision of deep learners (SCD-CD-DCNN) respectively.

Identifying key factors associated with commuting burden and modelling their non-linear relationships: the case study of Shenzhen, China

ABSTRACT Being closely related to quality of life, commuting burden has received increasing academic attention, especially for its influencing factors, such as land use and transportation infrastructures. However, scholars have seldom investigated different dimensions of commuting burden simultaneously. Furthermore, linear or generalised linear assumptions dominate previous studies regarding built environment-commuting burden associations. Therefore, employing mobile phone location data in Shenzhen, China, and utilising K-means Clustering and Gradient Boosting Decision Tree methods, this study examines the refined non-linear relationships between the built environment and commuting burden. It finds: (1) three commuting burden patterns are observed; (2) the built environment has the largest collective contribution for predicting commuting burden, and the residential environment is more important than the workplace environment; (3) nonlinearity generally exists in the relationships between the built environment and commuting burden. This study enriches the literature regarding commuting burden and its influencing factors and provides policy references for commuting burden mitigation.

Understanding and enhancing women’s loyalty in running events: a segmentation analysis based on brand perception

PurposeThis study examines factors that influence female runners’ commitment to running event brands, given the persistent disparity in participation rates between men and women in running events. Despite the increasing participation of women in popular races, there remains a gap in understanding the factors that drive their loyalty to these events. This research seeks to address this gap by analyzing female participants through segmentation based on profile data and marketing insights, with the aim of developing specific strategies to improve their loyalty.Design/methodology/approachA quantitative, cross-sectional study was conducted using a validated questionnaire administered to 383 women from two major running events. Hierarchical cluster analysis and K-means clustering were applied to segment participants based on brand-related variables (congruence, commitment, and trust), identifying two distinct groups: “Attached” and “Unattached” female runners.FindingsThe study identified two distinct groups of female runners, “attached” and “unattached,” based on their brand perceptions in marathon events. Attached runners scored higher on brand variables like congruence, commitment, and trust, and on performance variables such as satisfaction and word-of-mouth (WoM). Unattached runners, while newer to the sport, also valued congruence but had lower overall engagement. The findings suggest the need for tailored marketing strategies to enhance loyalty among both groups, focusing on different aspects like club memberships for attached runners and community-building efforts for unattached runners.Practical implicationsFor attached runners, event organizers should focus on loyalty programs, exclusive race experiences, and brand-driven community initiatives to reinforce engagement. Unattached runners require trust-building efforts, such as beginner-friendly programs, outreach initiatives, and customized entry-level engagement strategies, to encourage long-term participation. These insights are essential for sports marketers, race organizers, and policymakers seeking to enhance female inclusion and retention in endurance events.Originality/valueThis study is one of the first to apply brand perception segmentation to female runners, providing empirical evidence on the role of congruence, commitment, and trust in loyalty to sporting events. The findings contribute to the sport marketing literature by offering a data-driven approach to segmentation and event participation strategies, ensuring greater inclusion and long-term sustainability in running.

Advanced lightweight deep learning vision framework for efficient pavement damage identification

Pavement crack serves as a crucial indicator of road condition, directly associated with subsequent pavement deterioration. To address the demand for large-scale real-time pavement damage assessment, this study proposes a lightweight pavement damage detection model based on YOLOv5s (LPDD-YOLO). Initially, a lightweight feature extraction network, FasterNet, is adopted to reduce the number of parameters and computational complexity. Secondly, to mitigate the reduction in accuracy resulting from the usage of lightweight network, the attention-based downsampling module and the neural network cognitive module are introduced. These modules aim to enhance the feature extraction robustness and to eliminate interference from irrelevant features. In addition, considering the significant variation in aspect ratios and diverse morphologies of pavement damages, K-Means clustering and the deformable convolution module are employed. These mechanisms ensure dynamic anchor feature selection and extend the scope of deformation ability, respectively. According to the ablation experiment on a self-built dataset, LPDD-YOLO demonstrates notable improvements in both accuracy and efficiency compared to the original model. Specifically, the mAP increases by 4.1%, and the F1 score rises by 5.3%. Moreover, LPDD-YOLO can obtain a 47.3% reduction in parameters and a 54.4% decrease in GFLOPs. It is noteworthy that LPDD-YOLO achieves real-time and accurate damage detection, with a speed of up to 85 FPS. The effectiveness and superiority of LPDD-YOLO are further substantiated through comparisons with other state-of-the-art algorithms.

Association between uric acid to high-density lipoprotein cholesterol ratio and chronic kidney disease in Chinese patients with type 2 diabetes mellitus: a cross-sectional study

ObjectivesTo examine the association between uric acid (UA) to high-density lipoprotein cholesterol (HDL-C) ratio (UHR) and chronic kidney disease (CKD) in type 2 diabetes mellitus (T2DM) patients in China.MethodsThe investigation stems from a survey conducted in the eastern Chinese province of Zhejiang, spanning from March to November 2018. A multivariable logistic regression model was employed to assess the relationship between UHR and CKD, while restricted cubic spline (RCS) analysis was used to evaluate the dose–response relationship. Receiver operating characteristic (ROC) curve analysis was performed to determine the optimal UHR cut-off value and assess its diagnostic performance for CKD. Model performance was further evaluated using net reclassification improvement (NRI) and integrated discrimination improvement (IDI) metrics. Sensitivity analyses, including propensity score matching (PSM) and k-means clustering, were conducted to enhance the robustness of the findings. Subgroup analyses were performed across various demographic and clinical categories to examine the consistency of the UHR-CKD association.ResultsThis cross-sectional study included 1,756 Chinese patients with T2DM, among whom 485 (27.62%) were identified with CKD. Multivariable logistic regression analysis revealed a significant positive association between UHR and CKD. Per standard deviation (SD) increase in UHR was associated with a 40% higher odds of CKD (OR = 1.40, 95% CI: 1.23–1.60) after adjusting for potential covariates. When analyzed categorically, participants in the highest UHR tertile (T3) had 1.82-fold higher odds of CKD compared to the lowest tertile (T1) (95% CI: 1.32–2.50). RCS analysis demonstrated a consistent linear dose–response relationship between UHR and CKD across all models (all p for nonlinearity >0.05). ROC curve analysis identified an optimal UHR cut-off value of 12.28 for CKD prediction, with an area under the curve (AUC) of 0.710 (95% CI: 0.683–0.737) in the fully adjusted model. Subgroup analyses confirmed the robustness of the UHR-CKD association across most demographic and clinical variables, except for younger age groups (18–44 and 45–59 years) and smokers. Notably, BMI significantly modified the UHR-CKD relationship, with a nonlinear association observed in individuals with lower BMI (<24 kg/m2) and a linear association in those with higher BMI (≥24 kg/m2).ConclusionThis study demonstrates a significant dose–response relationship between the UHR and CKD in Chinese patients with T2DM, highlighting UHR as a promising biomarker for CKD risk assessment. The identified UHR cut-off of 12.28 offers a practical threshold for early renal monitoring and targeted interventions. Future research should explore UHR-targeted therapies and its integration into personalized risk stratification models to improve CKD management in T2DM.