Density Estimation Research Articles

Prioritization of Unknown LC-HRMS Features Based on Predicted Toxicity Categories.

Complex environmental samples contain a diverse array of known and unknown constituents. While liquid chromatography coupled with high-resolution mass spectrometry (LC-HRMS) nontargeted analysis (NTA) has emerged as an essential tool for the comprehensive study of such samples, the identification of individual constituents remains a significant challenge, primarily due to the vast number of detected features in each sample. To address this, prioritization strategies are frequently employed to narrow the focus to the most relevant features for further analysis. In this study, we developed a novel prioritization strategy that directly links fragmentation and chromatographic data to aquatic toxicity categories, bypassing the need for identification of individual compounds. Given that features are not always well-characterized through fragmentation, we created two models: (1) a Random Forest Classification (RFC) model, which classifies fish toxicity categories based on MS1, retention, and fragmentation data─expressed as cumulative neutral losses (CNLs)─when fragmentation information is available, and (2) a Kernel Density Estimation (KDE) model that relies solely on retention time and MS1 data when fragmentation is absent. Both models demonstrated accuracy comparable to that of structure-based prediction methods. We further tested the models on a pesticide mixture in a tea extract measured by LC-HRMS, where the CNL-based RFC model achieved 0.76 accuracy and the KDE model reached 0.61, showcasing their robust performance in real-world applications.

Are you ready for the tick season? Spring dynamic of tick diversity and density in urban and suburban areas

BackgroundOccurrence of tick-borne diseases (TBD) is often seasonal and associated with seasonal activity of appropriate tick vectors. As seasonal activity of ticks differs, the risk of contracting particular TBD should change between and within seasons. It is of key importance to monitor seasonal dynamic of tick vectors, especially in human-associated habitats. The aim of the current study was to compare activity and density of Ixodes ricinus and Dermacentor reticulatus during spring season in urban and suburban habitats.MethodsSystematic tick collection by dragging was performed every 1–2 weeks between mid-March and mid-June 2021 at 15 sites: 6 in Warsaw (urban areas) and 9 in suburban areas.ResultsDuring 178 field collections of ticks, including 131 collections from urban sites and 47 collections from rural areas, 738 ticks (385 adult D. reticulatus and 353 I. ricinus) were collected. Dermacentor reticulatus ticks are found from the beginning of spring, peaking in April and May, and I. ricinus ticks are present from early April, peaking in April and May as well. I. ricinus were abundant in rural and urban areas, including botanical garden and forest kindergarten area. Dermacentor reticulatus were found in urban fallow lands but were not collected in parks. These ticks were abundant in fallow lands, meadow, and mixed forest. DNA of B. burgdorferi s.l. and Rickettsia spp. was identified in ticks from urban areas.ConclusionsDue to the marked differences in spring dynamic of D. reticulatus and I. ricinus, the sampling effort should be repeated at least three times per season for accurate estimation of tick occurrence (presence/absence) and density. Due to “exchange” of tick species, total tick density remains high through the spring season of activity, which may result in high transmission of tick-borne pathogens (TBPs). Tick densities are dependent on the habitat type and may be low in well-managed agricultural habitats (crop fields, pastures, chicken yard), but high in semi-natural habitats (fallow lands, rural forests). Numerous I. ricinus populations can be maintained in urban green areas such as botanical gardens. Ticks from urban areas can serve as vectors of important TBPs (B. burgdorferi s.l., Rickettsia spp.).Graphical

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.

A study on the spatiotemporal evolutionary pattern and influencing factors of sports venues in Xi’an

Based on multi-source data, the spatio-temporal distribution characteristics, evolution trends, and driving factors of sports venues in Xi’an City were systematically explored in this work. Utilizing the Kernel Density Estimation (KDE) method, the distribution density of sports venues in 2010, 2015, and 2020 was scientifically calculated. Furthermore, the Multi-Scale Geographically Weighted Regression (MGWR) model was employed to analyse the factors influencing regional sports venues. Our analysis demonstrated that the number of sports venues in Xi’an has progressively increased, exhibiting a trend of gradually diffusing outwards from the centers of Xincheng District, Beilin District, and Lianhu District. Their spatial distribution is affected to varying degrees by commerce, industry, population, transportation, ecology, topography, and landforms, with the impact of commerce, industry, and population beginning to diminish. The transportation and environmental quality are gradually becoming core factors, and the influence of slope on sports venues shows spatial consistency. Moreover, the study also probes into the complexity of various influencing factors and provides a reference for the rational layout of sports venues in other cities. The importance of constructing and arranging sports venues judiciously in positively correlated areas is underscored to optimize their role in urban development.

The Coordinated Relationship Between the Tourism Economy System and the Tourism Governance System: Empirical Evidence from China

The development of tourism governance (TG) is influenced by the tourism economy (TE), and the development of TE is guaranteed by tourism governance. This study investigates the development levels of the tourism economy and tourism governance, as well as their interactive coordination in 31 Chinese provinces (including municipalities and autonomous regions) from 2012 to 2021. First, the vertical and horizontal differentiation method was employed to measure tourism economy and tourism governance development levels. Second, the Panel Vector Autoregression (PVAR) model was adopted to examine the Granger causality and the interactive effects between the tourism economy and tourism governance. Third, the coupled coordination model, kernel density estimation, and Markov chain model were combined to explore the degree of coordinated development and the spatio-temporal evolutionary trend of TE-TG. The findings reveal the following: (1) The development level of the tourism economy exhibits a fluctuating upward trend, with its spatial distribution pattern demonstrating a distinct coastal-to-inland decreasing gradient. Meanwhile, tourism governance shows a steady improvement trajectory marked by significant regional disparities. (2) A long-term equilibrium relationship has been established between the tourism economy and tourism governance, with bidirectional Granger causality observed between the two systems. (3) The coupled coordination between the tourism economy and tourism governance has progressively increased. However, the development level of tourism governance still lags behind that of the tourism economy. The eastern and central regions demonstrate significantly higher TE-TG coordination levels compared to the western and northeastern regions.

Design of a novel noise resilient algorithm for fault detection in wind turbines on supervisory control and data acquisition system

Wind turbine faults, including electrical, mechanical or aerodynamics-related, can potentially reduce operational efficiency, causing downtimes or, in some cases, leading to severe damage. Hence, timely detection of these operational anomalies is crucial for optimizing performance and reducing maintenance costs. The following study explores the application of Supervisory Control and Data Acquisition (SCADA) data for fault detection and diagnosing different operational states in wind turbines by focusing on environmental and operational factors which affect the performance. An efficient noise-resilient classification framework is proposed which includes a novel Perturbed-Random Forest (P-RF) algorithm to diagnose operational states with high accuracy even under noisy conditions. Seasonal discrepancies in power generation are analyzed using theoretical and actual power analysis. Further, distributions of features are realized via kernel density estimation and efficiency metrics in order to identify performance inefficiencies. The P-RF algorithm achieved 99.72% accuracy in diagnosing the status of the SCADA-based wind turbine although with a slightly higher computational complexity than a baseline Random Forest algorithm. Multiple Evaluation metrics were employed to assess the performance of the proposed model under different signal-to-noise conditions.

Spatial and Temporal Differentiation of Carbon Emission Efficiency and the Impact of Green Technology Innovation in Hubei Province

Green technological innovation is pivotal in advancing the ‘dual carbon’ target and promoting sustainable and low-carbon development. This research examines 17 prefecture-level cities in Hubei Province, employing the Super-SBM model for assessing emissions of carbon efficiency from 2010 to 2020. The kernel density estimation and the Dagum coefficient of Gini are also used to examine the spatio-temporal differentiation and the evolution of these efficiencies. A data panel regression model is utilized to evaluate how green technological innovation impacts carbon emission efficiency in Hubei Province. The research revealed that (1) Hubei Province’s carbon emission efficiency first fluctuated and then increased rapidly, and (2) the overall regional difference in carbon emission efficiency in Hubei Province shows a trend of first decreasing and then gradually increasing. The Wuhan metropolitan area and the Xiang-yang-Shiyan-Suizhou-Shennongjia urban area are quite different; the differentiation within the Yichang-Jingzhou-Jing-Enshi urban agglomeration shows a narrowing trend. (3) The innovation elements of green technology are positively correlated with the effectiveness of carbon emissions; the relationship between economic expansion and population density among the control variables also shows a positive correlation, while the industrial structure and government environmental regulations are negatively correlated. (4) In Hubei Province, there is a temporal lag between green technological innovation and its impact on carbon emission efficiency. Capital investment and technical achievement currently enhance carbon emission efficiency, while human capital positively affects carbon emission efficiency during a second lag period. This article proposes countermeasures and recommendations for R&D capital spending, innovative talent cultivation, and regional differentiation, providing specific references to advance the coordinated growth of the whole Hubei Province and green sustainable development.

BEACON: JWST NIRCam Pure-parallel Imaging Survey. I. Survey Design and Initial Results

Abstract We introduce the Bias-free Extragalactic Analysis for Cosmic Origins with NIRCam (BEACON) survey, a JWST Cycle 2 program allocated up to 600 pure-parallel hours of observations. BEACON explores high-latitude areas of the sky with JWST/NIRCam over ∼100 independent sight lines, totaling ∼0.3 deg2, reaching a median F444W depth of ≈28.2 AB mag (5σ). Based on existing JWST observations in legacy fields, we estimate that BEACON will photometrically identify 25–150 galaxies at z > 10 and 500–1000 at z ∼ 7–10 uniquely enabled by an efficient multiple filter configuration spanning 0.9–5.0 μm. The expected sample size of z > 10 galaxies will allow us to obtain robust number density estimates and to discriminate between different models of early star formation. In this paper, we present an overview of the survey design and initial results using the first 19 fields. We present 129 galaxy candidates at z ≳7 identified in those fields, including 11 galaxies at z ≳10 and several UV-luminous (M UV < −21 mag) galaxies at z ∼ 8. The number densities of z < 13 galaxies inferred from the initial fields are overall consistent with those in the literature. Despite reaching a considerably large volume (∼105 Mpc3), however, we find no galaxy candidates at z > 13, providing us with a complimentary insight into early galaxy evolution with minimal cosmic variance. We publish imaging and catalog data products for these initial fields. Upon survey completion, all BEACON data will be coherently processed and distributed to the community along with catalogs for redshift and other physical quantities.

Preventing Posterior Collapse with DVAE for Text Modeling

This paper introduces a novel variational autoencoder model termed DVAE to prevent posterior collapse in text modeling. DVAE employs a dual-path architecture within its decoder: path A and path B. Path A makes the direct input of text instances into the decoder, whereas path B replaces a subset of word tokens in the text instances with a generic unknown token before their input into the decoder. A stopping strategy is implemented, wherein both paths are concurrently active during the early phases of training. As the model progresses towards convergence, path B is removed. To further refine the performance, a KL weight dropout method is employed, which randomly sets certain dimensions of the KL weight to zero during the annealing process. DVAE compels the latent variables to encode more information about the input texts through path B and fully utilize the expressiveness of the decoder, as well as avoiding the local optimum when path B is active through path A and the stopping strategy. Furthermore, the KL weight dropout method augments the number of active units within the latent variables. Experimental results show the excellent performance of DVAE in density estimation, representation learning, and text generation.

A Slow and Underappreciated Forest Megafauna: Food Habits, Movements, and Multiscale Habitat Preferences of Critically Endangered Sundaic Giant Tortoises (Manouria emys emys).

Sundaic giant tortoises (Manouria emys emys) are the largest chelonians in Asia. Classified as critically endangered, they are extremely rare throughout their range. The limited knowledge of their behavior and ecology hampers effective conservation initiatives. We integrated GPS tracking, behavioral observations, local ecological knowledge, resource selection functions, spatial distribution modeling, and landscape functional connectivity to assess key aspects of their food habits, movement patterns, and habitat relationships at local and broad scale in Sumatra, Indonesia. Sundaic giant tortoises were predominantly diurnal (93% of activities) and had a mean home range of 27.5ha (±28.8 SD; N = 3 individuals, autocorrelated kernel density estimate). We identified 40 plant species from 20 families, including ferns, monocots, and dicots, consumed by the tortoises. They consumed the fruits of 30 of these plants, swallowing seeds as large as those of Durio and Artocarpus (>2 × 4cm), indicating potential seed dispersal over distances exceeding 1km. Habitat preference analysis at both local and island-wide scales showed a consistent preference for primary rainforests with steep slopes at moderate altitudes, and near rivers, mostly on the western side of the island. We identified 15 "priority areas," that is, those with high habitat suitability but lacking formal protection. Among these, only one, Gunung Talang, lacked connectivity with other forest areas, prompting us to propose a wildlife corridor connecting it to Kerinci Seblat National Park. In addition, we argue that Sundaic giant tortoises hold potential for rewilding operations in Sumatra and other parts of their range, emphasizing the need for targeted conservation efforts.

Study on the Spatial Distribution Patterns and Driving Forces of Rainstorm-Induced Flash Flood in the Yarlung Tsangpo River Basin

Flash floods, typically triggered by natural events such as heavy rainfall, snowmelt, and dam failures, are characterized by abrupt onset, destructive power, unpredictability, and challenges in mitigation. This study investigates the spatial distribution patterns and driving mechanisms of rainstorm-induced flash flood disasters in the Yarlung Tsangpo River Basin (YTRB) by integrating topography, hydrometeorology, human activity data, and historical disaster records. Through a multi-method spatial analysis framework—including kernel density estimation, standard deviation ellipse, spatial autocorrelation (Moran’s I and Getis–Ord Gi*), and the optimal parameter geographic detector (OPGD) model (integrating univariate analysis and interaction detection)—we reveal multiscale disaster dynamics across county, township, and small catchment levels. Key findings indicate that finer spatial resolution (e.g., small catchment scale) enhances precision when identifying high-risk zones. Temporally, the number of rainstorm-induced flash floods increased significantly and disaster-affected areas expanded significantly from the 1980s to the 2010s, with a peak spatial dispersion observed during 2010–2019, reflecting a westward shift in disaster distribution. Spatial aggregation of flash floods persisted throughout the study period, concentrated in the central basin. Village density (TD) was identified as the predominant human activity factor, exhibiting nonlinear amplification through interactions with short-duration heavy rainfall (particularly 3 h [P3] and 6 h [P6] maximum precipitations) and GDP. These precipitation durations demonstrated compounding risk effects, where sustained rainfall intensity progressively heightened disaster potential. Topographic and ecological interactions, particularly between elevation (DEM) and vegetation type (VT), further modulate disaster intensity. These findings provide critical insights for risk zonation and targeted prevention strategies in high-altitude river basins.

Life stage and seasonal habitat use of the porbeagle Lamna nasus in the Northwest Atlantic Ocean

For highly mobile marine species such as pelagic elasmobranchs, the development of effective spatial management requires a comprehensive understanding of movement ecology. Research incorporating movement data across seasons and life stages, including reproductive states, is valuable for informing spatial management, yet is absent for most species. In the Northwest (NW) Atlantic Ocean (hereafter referred to as NW Atlantic), the porbeagle shark Lamna nasus is a pelagic species that is overfished, has a live retention ban (Canada) or landings regulations (United States), and is also commonly captured incidentally as bycatch. Research on the spatiotemporal dynamics of NW Atlantic porbeagle habitat use is limited, with all previous research utilizing pop-off satellite archival tags that are prone to large uncertainty in location estimates. This study used higher-accuracy fin-mount satellite tags to identify patterns in habitat use across life stages and seasons for porbeagle sharks tagged off the northeastern coast of the United States. During the summer and fall, the 95% kernel density estimate (referred to as “activity space”) of tagged porbeagles occurred almost exclusively on the continental shelf in the Gulf of Maine and Georges Bank. Activity space expanded and shifted southwards to include offshore environments during the winter and spring for juveniles, mature non-gravid females, and mature females of unknown reproductive states, while the activity space of mature males and gravid females remained in shelf waters year-round. This finding differs from the previous assumption that southward migrations are linked to reproduction for NW Atlantic porbeagles. Tagged porbeagles were also found to have a relatively small 50% kernel density estimate (referred to as “high occupancy area”) located around Cape Cod, Massachusetts that was well-conserved across life stages and seasons. This relatively static, small high occupancy area has implications for the population’s conservation given the high amount of fishing activity (rod-and-reel, trawl, gillnet) occurring within this region. Given the overlap between porbeagle high occupancy area and fishing activity, as well as the relatively high recapture rate of tagged sharks (10.5%), the coastal waters around Cape Cod, Massachusetts should be considered for spatial management of the NW Atlantic porbeagle.

A Novel Approach to Evaluating Industrial Spatial Structure Upgrading: Evidence from 284 Cities and 96 Sub-Industries in China (1978–2022)

This paper introduces a novel spatial angle index method for evaluating industrial spatial structure upgrading, drawing on a review and critique of existing approaches, including the traditional proportional method, industrial structure hierarchy coefficient, the production efficiency method and the cosine angle method. The proposed method integrates both local and projected spatial angle indices, accounting for the actual industry proportions of specific regions and assessing the deviation of these indices from extreme states. This approach effectively captures spatial spillover effects between regions and the combined influence of local and projected points on industrial spatial structure upgrading. Utilizing firm survival big data from 96 sub-industries across 284 cities in China from 1978 to 2022, the paper evaluates the upgrading levels of industrial spatial structure in these cities and examines their spatiotemporal evolution patterns using kernel density estimation. The study reveals that although different spatial weight matrices (commuting distance, latitude–longitude distance, and commuting time) produce slightly different results, the differences are not substantial. Notably, the analysis shows that the tertiary sector consistently demonstrates superior upgrading levels, while the secondary sector has underperformed, particularly since 1992. The primary sector, however, has experienced significant improvements, at times even surpassing the tertiary sector. The findings further suggest that while significant changes in industrial spatial structure upgrading occurred before 1987, the pace of change has stabilized since 1988.

Artificial intelligence approach for estimating energy density of liquid metal batteries

Achieving a high energy density in liquid metal batteries (LMBs) still remains a big challenge. Due to the multitude of affecting parameters within the system, traditional ways may not fully capture the complexity of LMBs. The artificial intelligence approach can be effectively applied to deal with low energy density issues. Herein, we represented the first implementation of the Gaussian Process Regression to predict the LMBs’ energy density to attain the highest accuracy compared to existing models. Four different kernels, namely Exponential, Matern5/2, Rational Quadratic, and Squared Exponential were utilized to achieve the most accurate GPR model. A huge dataset containing 2158 LMB datapoint was gathered from the literature. It contains 41 input parameters, including alloy-related, LMB-related, and creative features. The GPR-Exponential model showed the greatest battery energy density estimate accuracy among the proposed models. The training and testing R2 values were 0.9976 and 0.9975, respectively, indicating the near-perfect accuracy which makes it the most precise model that has been presented so far. According to sensitivity analysis outcomes, it can be claimed that Sb mole fraction, average ionization energy, and average melting temperature with the respective relevancy factors of 0.6672, 0.6550, and 0.6507 could noticeably affect the LMBs’ energy density. Furthermore, the results showed that the LMBs’ energy density is more sensitive to the electrode-dependent and operational parameters rather than the electrolyte situation.

Towards Better Spherical Sliced-Wasserstein Distance Learning with Data-Adaptive Discriminative Projection Direction

Spherical Sliced-Wasserstein (SSW) has recently been proposed to measure the discrepancy between spherical data distributions in various fields, such as geology, medical domains, computer vision, and deep representation learning. However, in the original SSW, all projection directions are treated equally, which is too idealistic and cannot accurately reflect the importance of different projection directions for various data distributions. To address this issue, we propose a novel data-adaptive Discriminative Spherical Sliced-Wasserstein (DSSW) distance, which utilizes a projected energy function to determine the discriminative projection direction for SSW. In our new DSSW, we introduce two types of projected energy functions to generate the weights for projection directions with complete theoretical guarantees. The first type employs a non-parametric deterministic function that transforms the projected Wasserstein distance into its corresponding weight in each projection direction. This improves the performance of the original SSW distance with negligible additional computational overhead. The second type utilizes a neural network-induced function that learns the projection direction weight through a parameterized neural network based on data projections. This further enhances the performance of the original SSW distance with less extra computational overhead. Finally, we evaluate the performance of our proposed DSSW by comparing it with several state-of-the-art methods across a variety of machine learning tasks, including gradient flows, density estimation on real earth data, and self-supervised learning.

A probabilistic sampling strategy for estimating plant density in Posidonia oceanica meadows

Marine and coastal ecosystems, such as seagrasses, mangroves, and coral reefs, provide a range of essential provisioning, regulating and cultural ecosystem services. Recent United Nations guidelines on ecosystem accounting (SEEA EA) emphasise the need for biophysical data as the foundation for compiling ecosystem accounts and conducting economic evaluations for developing indicators and informing policies and interventions. However, data availability on marine ecosystems is limited with respect to terrestrial ones. Moreover, the collection of biophysical data on marine ecosystem extent and condition required for ecosystem accounting (EA) is often not aligned with existing habitat monitoring strategies. This study aims to address the scarcity of spatial data on marine ecosystems and facilitate the integration of current monitoring strategies with the scope of EA. We propose the application of design-based inference for the estimation, mapping, and monitoring of key ecological attributes of marine ecosystems. We focus on the habitat of Posidonia oceanica, an endemic seagrass of the Mediterranean Sea, but the proposed strategy is adaptable to other ecosystems. The benefits of appropriate probabilistic sampling schemes for assessing P. oceanica are explored via simulation testing. The performance of different sample schemes in artificial populations reveals that reliable estimates of density (as well as their precision) can be obtained even with low sample sizes. The empirical viability of our methodology is exemplified using data collected on a meadow located in an Italian Marine Protected Area (Puglia region, Southern Italy).

Hypergraph Attacks via Injecting Homogeneous Nodes into Elite Hyperedges

Recent studies have shown that Hypergraph Neural Networks (HGNNs) are vulnerable to adversarial attacks. Existing approaches focus on hypergraph modification attacks guided by gradients, overlooking node spanning in the hypergraph and the group identity of hyperedges, thereby resulting in limited attack performance and detectable attacks. In this manuscript, we present a novel framework, i.e., Hypergraph Attacks via Injecting Homogeneous Nodes into Elite Hyperedges (IE-Attack), to tackle these challenges. Initially, utilizing the node spanning in the hypergraph, we propose the elite hyperedges sampler to identify hyperedges to be injected. Subsequently, a node generator utilizing Kernel Density Estimation (KDE) is proposed to generate the homogeneous node with the group identity of hyperedges. Finally, by injecting the homogeneous node into elite hyperedges, IE-Attack improves the attack performance and enhances the imperceptibility of attacks. Extensive experiments are conducted on five authentic datasets to validate the effectiveness of IE-Attack and the corresponding superiority to state-of-the-art methods.

A Unifying Information-theoretic Perspective on Evaluating Generative Models

Considering the difficulty of interpreting generative model output, there is significant current research focused on determining meaningful evaluation metrics. Several recent approaches utilize "precision" and "recall," borrowed from the classification domain, to individually quantify the output fidelity (realism) and output diversity (representation of the real data variation), respectively. With the increase in metric proposals, there is a need for a unifying perspective, allowing for easier comparison and clearer explanation of their benefits and drawbacks. To this end, we unify a class of kth-nearest neighbors (kNN)-based metrics under an information-theoretic lens using approaches from kNN density estimation. Additionally, we propose a tri-dimensional metric composed of Precision Cross-Entropy (PCE), Recall Cross-Entropy (RCE), and Recall Entropy (RE), which separately measure fidelity and two distinct aspects of diversity, inter- and intra-class. Our domain-agnostic metric, derived from the information-theoretic concepts of entropy and cross-entropy, can be dissected for both sample- and mode-level analysis. Our detailed experimental results demonstrate the sensitivity of our metric components to their respective qualities and reveal undesirable behaviors of other metrics.

Spatial and breeding site analysis of Aedes spp. at dengue-prone areas in Kuala Lumpur, Malaysia

ABSTRACT Understanding Aedes breeding sites is crucial for effective vector control, yet the application of Geographic Information Systems (GIS) and spatial analysis in Kuala Lumpur is limited. This study explores the spatial distribution and characteristics of Aedes aegypti and Aedes albopictus breeding sites in dengue-prone areas of Bandar Tasik Selatan. Entomological data from the Kuala Lumpur Health Department (2022–2023) were analyzed using Average Nearest Neighbor (ANN), Global Moran’s I, and Kernel Density Estimation (KDE). Logistic regression and principal component analyses (PCA) were used to assess breeding site characteristics. Of 6,027 water containers inspected, 402 (6.7%) were positive for Aedes larvae. ANN and Moran’s I analyses revealed significant clustering in Zone A and Zone B, while KDE identified breeding hotspots. Logistic regression indicated higher risk in plastic (aOR = 69.58), rubber (aOR = 35.11), and cement (aOR = 24.70) containers, while rainwater (aOR = 0.24), tap water (aOR = 0.27), and partial shading (aOR = 0.41) reduced risk. PCA revealed key breeding site variations across residential, public, school, and food areas. These findings support targeted vector control, efficient resource allocation, and tailored interventions to mitigate dengue risks in high-priority areas.

SatCLIP: Global, General-Purpose Location Embeddings with Satellite Imagery

Geographic information is essential for modeling tasks in fields ranging from ecology to epidemiology. However, extracting relevant location characteristics for a given task can be challenging, often requiring expensive data fusion or distillation from massive global imagery datasets. To address this challenge, we introduce Satellite Contrastive Location-Image Pretraining (SatCLIP). This global, general-purpose geographic location encoder learns an implicit representation of locations by matching CNN and ViT inferred visual patterns of openly available satellite imagery with their geographic coordinates. The resulting SatCLIP location encoder efficiently summarizes the characteristics of any given location for convenient use in downstream tasks. In our experiments, we use SatCLIP embeddings to improve performance on nine diverse geospatial prediction tasks including temperature prediction, animal recognition, and population density estimation. Across tasks, SatCLIP consistently outperforms alternative location encoders and shows promise for improving geographic domain adaptation. These results demonstrate the potential of vision-location models to learn meaningful representations of our planet from the vast, varied, and largely untapped modalities of geospatial data.