Search Results (4,330)

This article presents the use of autoencoders as part of a dimensionality-reduction method in the task of crime clustering. The study was conducted on a real-world crime dataset from the city of Chicago, based on publicly available police records. The initial data processing involved selecting and extracting variables, aggregating the selected variables, and converting crime categories and incident locations into contextual embeddings. The data prepared in this way was used to train various autoencoder architectures, including Vanilla, convolutional, denoising and variational models. The representations obtained from the latent layer of the encoder were then used as input data for clustering methods, such as k-means, Gaussian mixture model, and spectral clustering. The experimental results showed that the use of autoencoders in the clustering process enabled the identification of distinct groups of offences, with the best results (measured using the ARI and NMI metrics) obtained for Vanilla autoencoders combined with k-means and GMM, particularly with intermediate latent-space dimensions. The results confirm the potential of autoencoders as effective tools for dimensionality reduction and feature extraction in crime data analysis, as well as their usefulness in the exploratory analysis of complex urban data. Full article

Salt stress severely restricts soybean seedling growth and yield formation, and the redundant indicators and low efficiency of conventional salt tolerance evaluation methods limit the large-scale screening and breeding of salt-tolerant soybean germplasms. In this study, we aimed to establish a simplified and efficient salt tolerance evaluation system for soybean seedlings under saline pond conditions. We determined 16 phenotypic traits of 100 soybean germplasm resources under soil salt stress (0.3% soil salt content, EC 5.0 dS/m), calculated the salt tolerance coefficient (STC) of each trait, and comprehensively analyzed phenotypic variation, correlation, germplasm classification, and core evaluation indices via principal component analysis (PCA), K-means clustering, random forest model, SHAP interpretation, 10-fold nested cross-validation, and correlation network analysis. Biomass-related traits exhibited abundant phenotypic variation, with coefficients of variation ranging from 42.6% to 48.4%. Five principal components explained 82.90% of the total phenotypic variation and divided the accessions into four salt tolerance categories. Total fresh weight (TFW), stem fresh weight (SFW), and leaf fresh weight (LFW) were identified as the core indices, together accounting for over 93% of the total feature importance in the random forest model, whereas the remaining 13 traits each contributed less than 1.2%. The simplified three-index model showed strong consistency with the full 16-trait model (Pearson r > 0.970, AUC = 0.970) and achieved a screening accuracy of 90.0% under 10-fold nested cross-validation. Under the experimental conditions examined, fresh biomass accumulation emerged as the dominant phenotypic characteristic associated with seedling salt tolerance. This simplified evaluation framework may facilitate rapid preliminary screening of salt-tolerant soybean germplasms at the seedling stage, pending further validation across diverse environments and genetic backgrounds. Full article

Green exposure in public spaces plays a critical role in promoting mental health and well-being, especially for populations experiencing elevated psychological pressure in academically demanding environments such as university campuses. However, existing research remains largely limited to coarse-scale assessments, often overlooking the nuanced and nonlinear interactions between visual environmental characteristics and restorative quality across diverse spatial typologies. Addressing this gap, this study proposes a fine-scale, data-driven framework that integrates computer vision (CV) with perception-based evaluations using Shandong Jianzhu University as a case study. A semantic segmentation model (Mask2Former) was employed to extract 14 visual environmental elements from 193 campus photographs, while 426 students provided 4260 image-based evaluations of Perceived Restorative Quality (PRQ). Ridge regression optimized through cross-validation was used to identify key environmental correlates of PRQ, K-means clustering was applied to classify Campus Restorative Space (CRS) typologies, and Generalized Propensity Score Matching (GPSM) was employed to examine nonlinear dose–response relationships between sensory dimensions and restorative quality. The results indicate that natural elements, particularly vegetation and water-related features, are generally associated with higher PRQ scores; however, restorative responses do not consistently follow linear patterns. Several sensory dimensions exhibited threshold effects and diminishing marginal returns, suggesting that restorative benefits depend on balanced environmental composition rather than the simple accumulation of green elements. Furthermore, three CRS typologies—Greenery-Dominant Spaces, Open Landscape Spaces, and Courtyard Spaces—demonstrated distinct restorative pathways, indicating that the effects of environmental characteristics vary across spatial contexts. These findings support a context-dependent understanding of restorative environments and highlight the importance of spatial typology, environmental composition, and nonlinear responses in shaping restorative experiences. The study proposes a transferable analytical framework for assessing campus restorative environments and offers practical implications for evidence-based campus planning and design. Full article

Traditional soil management relies on “whole-field” averages, which leads to resource waste and environmental degradation under anthropogenic pressures. While combining electromagnetic induction (EMI) and remote sensing is known for digital soil mapping, current approaches lack a unified, automated framework to handle complex multi-source data dependencies for local-scale precision irrigation. To overcome this limitation, this study introduces a novel integrated methodology that couples high-resolution geophysical datasets and remote/proximal sensing through an automated machine learning workflow, capturing dynamic soil–human interaction boundaries more precisely than traditional empirical overlays. The general methodology was tested in a vineyard plot within the Torre Guaceto Natural Reserve (Southern Italy). Spatial datasets from EMI and remote sensing were integrated. Crucially, the K-means clustering algorithm was deployed early in the workflow to optimize the fused datasets and classify the plot into homogeneous zone clusters. The machine learning approach successfully identified two distinct main soil clusters. The spatial boundaries of these zones were rigorously validated using in situ soil moisture data from capacitance sensors, showing a statistically significant variance in volumetric water content between the two zones. This study demonstrates that integrated machine learning workflows can accurately delineate precision agricultural zones without relying on high-cost exhaustive sampling. It is recommended that farmers and managers within sensitive nature reserves adopt this cluster-based Variable Rate Application (VRA) for water and fertilizers to optimize resource efficiency and prevent nutrient leaching into underlying aquifers. Full article

The accurate measurement of solid concentration in horizontal gas–solid flows is very important to guarantee production efficiency and process control. However, gravity causes uneven particle distribution, which creates a strong nonlinear relationship between sensor signals and solid concentration. When particle distribution changes, a single linear measurement model cannot provide enough detection accuracy. This paper proposes an adaptive model matching strategy for solid concentration measurement in horizontal gas–solid flows by using array capacitive sensor. The array capacitive sensor works with two excitation modes. The concave-ring excitation mode collects particle distribution information, and the multi-electrode excitation mode obtains solid concentration. These two types of signals build a dynamic matching relationship between particle distribution features and measurement models. In detail, signals from concave-ring excitation are input into the BP-Adaboost algorithm to classify particle distribution states. Multiple linear measurement models between multi-electrode signals and solid concentration are built through K-means clustering. After recognizing the particle distribution type, the Euclidean distance is used to automatically select the corresponding measurement model. A 3D simulation model combining gas–solid two-phase flow and electrostatic field is set up to test the feasibility of the proposed measurement method. Laboratory experiments are also conducted to prove its reliability. The test results show that the method can adapt well to various particle distribution states. Within the solid concentration range of 0.34–14.08%, the average relative measurement error is 4.41%. This method effectively improves the measurement accuracy of solid concentration in horizontal gas–solid two-phase flow. Full article

Objectives: The analysis of basketball performance has increasingly incorporated advanced analytics and machine learning methods to better understand the factors that influence offensive efficiency and match dynamics. The present study aimed to identify performance profiles in basketball using unsupervised machine learning techniques and to examine the physical load and performance indicators that differentiate these profiles. Methods: Team-quarter observations from the Final 8 phase of the Greek U16 Basketball Championship were stratified into quarters with large score differences and quarters with small score differences according to the quarter-specific score differential and the sample median of 5 points. K-means clustering was applied separately to each dataset to identify latent performance patterns. Candidate solutions were evaluated using the Elbow method, Silhouette coefficient, Calinski–Harabasz Index, and Davies–Bouldin Index. Based on their combined interpretation, together with considerations of parsimony and practical interpretability, two-cluster solutions were retained for both datasets. Cluster stability was assessed using the Adjusted Rand Index (ARI), while t-distributed stochastic neighbor embedding (t-SNE) was used exclusively for visualization of the identified clusters. Results: Welch’s independent-samples t-tests with Benjamini–Hochberg false discovery rate (FDR) correction identified significant differences between clusters across several external load variables, including jump load, total distance covered, accumulated acceleration load, and distance covered in different speed zones (pFDR < 0.001). Clusters characterized by higher movement intensity also exhibited higher values for basketball performance and offensive-efficiency indicators. Although higher-performance clusters showed numerically higher winning proportions in both contexts (large score differences: 70.0% vs. 45.7%; small score differences: 56.7% vs. 40.6%), chi-square analyses indicated that cluster membership was not significantly associated with quarter outcomes. Conclusions: Overall, the findings suggest that performance profiles in basketball are primarily differentiated by external-load characteristics, particularly movement intensity, and offensive-performance indicators, highlighting the importance of integrating both physical and technical performance indicators in basketball performance analysis. Full article

The wide integration of volatile renewable energy has intensified voltage regulation challenges in distribution networks. Conventional centralized control approaches are inherently limited by slow response times, heavy reliance on communication, and poor coordination. To address this issue, this paper proposes a voltage optimization method that integrates adaptive partitioning and hierarchical coordination. First, node electrical coupling characteristics are extracted based on the voltage-reactive power sensitivity matrix analysis. The elbow method and K-means clustering are jointly applied to achieve network partitioning, where the optimal number of partitions is automatically determined. On this basis, a hierarchical coordination control architecture is established. The upper level employs scenario-based worst-case dispatch to minimize comprehensive operating costs and voltage deviations, generating reactive power regulation references for distributed generators in each partition. The lower level uses an improved droop control based on the voltage-squared relationship, enabling distributed generators to adjust reactive power in real time according to local voltage deviations, thereby achieving coordination between global optimization and local response. Lastly, simulation results on a modified IEEE 33-bus system demonstrate the effectiveness of the proposed method in eliminating voltage violations, particularly under heavy-load conditions. Full article

With the increasing variety of electric vehicles (EVs) joining virtual power plants (VPPs), VPP operators increasingly require precise and tailored models for schedulable EV energy. Based on a publicly available anonymous EV charging power dataset, EV users are clustered through a weighted K-means++ algorithm. Secondly, based on the results of clustering, we analyzed the daily traveling patterns of various types of EVs, including commuting EVs, electric light-duty trucks (ELDTs) and electric tractors (ETs), and then customized the all-day schedulable energy domain model (SEDM) for each category. Subsequently, an optimal bidding strategy for a VPP consisting of diversified-member EVs, air conditionings (ACs), energy storage (ES) and distributed energy resources (DERs) is constructed. By modifying the levels of participation in supplementation and absorption of DERs among VPP members, while integrating considerations such as user comfort, EV defying rate, and seasonal variability, diverse VPP operational frameworks are established. Finally, using the Gurobi solver, the optimal bidding strategies and profit results under different scenarios are derived. The results indicate that (1) increasing the VPP members’ participation in the supplementation and absorption of DERs will bring higher benefits to both the VPP and its members; (2) with the increased sensitivity of users to room temperature and range anxiety, the demand response capacity of AC clusters decreases, reducing EV clusters’ market participation and VPP profits; and (3) among various types of EVs, ELDTs and ETs have a larger battery energy adjustment range, which can fully supplement the output shortfalls of DERs. Therefore, these EVs prove to be a good supplement for the improvement of VPP’s schedule capability and profitability. Full article

Academic competitions have become increasingly visible in higher education. This study examines the recorded expansion and distribution of computer science competition awards across Chinese universities. It treats these records as evidence of an award-visible university competition system with potential co-curricular functions, not as direct measures of curricular integration, student learning, institutional strategy, or internal resource flows. The conceptual framework distinguishes direct observations, derived descriptive indicators, and untested feedback propositions. The dataset contains national award records from 2012 to 2025. The main analysis uses complete annual data from 2012 to 2024; the 2025 records are used only for a provisional continuity check. Award-visible universities increased from 226 in 2012 to 1110 in 2024, and annual award records rose from fewer than 2000 to more than 40,000. Recorded awards also became less concentrated across competition categories. Inter-university inequality remained high within the award-visible sample, and the selected Theil decomposition attributed 91.47% of measured inequality to the within-province component. Adjacent-year rank correlations indicated positional stability among universities active in both years. K-means clustering identified four descriptive portfolio configurations. The results describe expansion, diversification, inequality, persistence, and portfolio heterogeneity within the recorded award system but do not establish the organizational mechanisms represented in the causal-loop framework. Full article

Overseas Chinese Yanglou dwellings in Southern Fujian are important architectural heritage combining Western architectural elements and local construction traditions. However, their diverse facade forms and incomplete records make manual classification inefficient and difficult to quantify. Taking Jinjiang, China, as the study area, this research constructs a Yanglou facade image dataset and proposes a deep learning-based workflow for type recognition, model interpretation, and spatial analysis. Deep visual features and unsupervised clustering were used to identify three morphological patterns: traditional continuity, partial addition, and overall transformation. These patterns were further organized into a classification system of three categories and seven subtypes. YOLOv8n-cls, YOLO11n-cls, and YOLO26n-cls were compared for facade recognition, while feature map visualization, Grad-CAM, and t-SNE were used for model interpretation. GIS analysis suggests that traditional-continuity types are mainly distributed in inland plain and piedmont settlements. The Five-Foot Way type is widely distributed, whereas overall-transformation types are concentrated in several coastal or near-coastal villages. YOLO11n-cls achieved the best overall validation performance within the current Jinjiang sample set. The proposed workflow supports quantitative classification, interpretable recognition, and spatial documentation of Yanglou dwellings. Full article

Recent studies indicate that regional disparities and differences in policy effectiveness have increased the importance of behavioural factors, including awareness, social norms, motivation, and cultural practices, in shaping energy transition outcomes. This study adopts an interdisciplinary framework that integrates behavioural, socio-technical, and regional development perspectives to explain household energy consumption behaviour and to translate behavioural evidence into policy recommendations. The study aims to identify and empirically assess the profiles of household energy consumption behaviour during the energy transition, examine how these profiles shape behavioural patterns, and develop differentiated policy recommendations. The empirical analysis is based on a survey of 997 household representatives from different regions of Ukraine. Exploratory factor analysis was applied to identify the main behavioural dimensions, K-means cluster analysis was used to classify respondents into behavioural profiles, and Pearson’s χ² test was employed to examine the association between categorical characteristics and cluster membership. The findings reveal three distinct behavioural clusters characterized by different levels of awareness, engagement, and readiness to adopt energy-efficient practices. The identified clusters represent individual-level behavioural profiles rather than territorial categories. Although the region of residence was significantly associated with cluster membership, regional characteristics should be interpreted as associated with the socio-economic, climatic, and infrastructural context in which different behavioural profiles emerge rather than determining individual behaviour directly. The findings provide evidence for integrating behavioural segmentation with regional characteristics when designing energy efficiency and demand-side management policies, thereby supporting more targeted and effective energy transition strategies in Ukraine. Full article

Family farming cooperatives and associations play a key role in rural development but often face managerial constraints that limit their organizational performance and access to institutional markets. This study assessed short-term changes in organizational maturity among organizations participating in Brazil’s Mais Gestão Program. A quasi-experimental pre–post design without a control group was adopted using paired observations from 42 organizations assessed at baseline (April 2025) and follow-up (November 2025). Organizational maturity was measured using the official Mais Gestão diagnostic instrument, which comprises six managerial dimensions. Descriptive statistics, paired-samples t-tests, Principal Component Analysis (PCA), ANOVA, and K-means Cluster Analysis were performed. Overall organizational maturity improved between assessments, particularly in governance, people management, financial management, commercial management, and production processes. PCA showed that the first principal component explained 65.48% of the total variance, supporting the view of organizational maturity as a multidimensional construct. These findings suggest that institutional support initiatives may strengthen managerial capacities and improve organizational readiness to access public policy opportunities. Full article

Background: Men who have sex with men (MSM) experience a disproportionate burden of human papillomavirus (HPV)-related disease, yet vaccination uptake remains uneven. With MSM as the key population of interest, this review synthesized evidence on uptake, willingness, multilevel barriers and opportunities, and vaccine-oriented outcomes among MSM and related sexual and gender minority populations when MSM-relevant findings could be extracted. Methods: Six databases were searched for original English- or Chinese-language studies published from 1 January 2010 to 31 December 2025. Findings were synthesized narratively. Barriers and opportunities were mapped using a five-level socio-ecological model; study-level willingness–uptake patterns were explored using K-means clustering as an exploratory analysis among the nine studies reporting both outcomes; and genotype-specific and immunogenicity findings were summarized separately. Results: Fifty-three studies involving 169,241 participants were included; 87.2% of participants were MSM, and 79% of studies were cross-sectional. Barriers and opportunities occurred across individual, provider-related interpersonal, organizational/institutional, community, and policy/societal levels. Recurrent paired mechanisms involved knowledge and trust, provider recommendation and communication, service accessibility, community support, and eligibility and affordability. Nine studies contributed data to the exploratory clustering analysis; seven fell into low-uptake clusters, including three with high willingness but low uptake, which may indicate a possible study-level implementation gap. Two studies reporting genotype prevalence and antibody responses provided limited biological context supporting vaccination before exposure and suggesting potential benefit from catch-up vaccination. Discussion: Among MSM and related sexual and gender minority populations, HPV vaccination is shaped by interacting individual and structural conditions, with MSM remaining the principal focus of the evidence base. Improving uptake is likely to require complementary strategies: universal, gender-neutral routine vaccination in adolescence before their sexual debut, and targeted catch-up for MSM who missed routine vaccination, supported by trusted provider endorsement, convenient service delivery, community engagement, and inclusive, affordable policy. The thematic counts were study-level and unweighted, and the nine-study clustering analysis was exploratory rather than definitive. Full article

Rapid urbanization has intensified urban heat issues. Previous studies often relied on subjective block selection and rarely integrated vegetation data. This study extracted vegetation from Wuhan’s satellite imagery and combined it with building geometry to generate large-scale 3D block models. Typical blocks were identified by clustering, and thermal environments were simulated using ENVI-met to establish regression models. POI and spatial analyses validated the results. The study found that 1. t-SNE outperforms PCA and UMAP in dimensionality reduction. 2. K-means surpasses GMM, DBSCAN, and Spectral in clustering. 3. ${SVF}_{ave}$, ${FA}_{all}$, ${VD}_W$, $VAR$, and BBA are critical for block morphology classification and block outdoor thermal assessment. 4. The final ridge regression model based on these indices achieved high R² values (0.805, 0.507, and 0.855), indicating excellent model performance. 5. The blocks in Cluster 1 (west of the Yangtze River) exhibit higher mean air temperatures. 6. the blocks in Cluster 2 (new areas) have high vegetation coverage, causing larger temperature differences between the inside and outside of blocks. This study provides a comprehensive workflow for urban block morphology classification and thermal assessment. Full article

The conservation of historic districts is often evaluated on the basis of visible material features, yet commercialized renewal may weaken visual and auditory cues of locality while preserving building façades and street textures. We established 340 sampling points at 20 m intervals along Ximen Street Historic District in Qujing, Yunnan Province, and constructed a streetscape–audio matched sample dataset. Through multimodal model-assisted recognition, K-means clustering, XGBoost, and SHAP methods, we identified the degree of visual and auditory homogenization and analyzed the nonlinear effects and interactions of visual and auditory factors on the perception of homogenization. The results show that (1) visual homogenization presents a continuous distribution, while soundscape homogenization presents a patch-like distribution and is higher overall; (2) commercial symbolism, generic decoration, and open interfaces lead to visual substitutability, while commercial broadcasting and mechanical sounds intensify soundscape standardization; dialects, residents’ conversations, and diverse sound events enhance the perception of locality; and (3) visual commercialization and soundscape standardization have a clear interaction effect. This case study provides insights into the living conservation and renewal of historic districts in similar contexts and may offer a fine-grained analytical reference for collaborative governance that considers both visible townscape features and daily soundscape locality. Full article