Spatial Dependence Research Articles

Design of an Iterative Method for Enhanced Retinal Image Analysis Using Stacked Deep Learning Operations

The necessity for advanced diagnostic techniques in ophthalmology has become increasingly evident, particularly for conditions such as glaucoma and diabetic retinopathy, which necessitate precise retinal image analysis. Current methodologies, while effective to a degree, fall short in terms of accuracy, efficiency, and adaptability, especially in semantic segmentation and disease classification tasks. These limitations underscore the need for more sophisticated and reliable approaches to retinal image analysis. In response, this work introduces a comprehensive framework leveraging Fully Convolutional Neural Networks (FCNNs) for optic disc segmentation and Cup-to-Disc Ratio (CDR) estimation. The selection of FCNNs is predicated on their demonstrated proficiency in semantic segmentation, ability to discern spatial dependencies, and capacity to learn intricate structures within fundus images without relying on manually engineered features. This approach aims to surpass existing segmentation accuracy benchmarks, targeting over 95% Intersection over Union (IoU) while reducing the mean absolute error in CDR estimation to below 10%. To address the challenges of data variability and quality, the framework incorporates Contrast Limited Adaptive Histogram Equalization (CLAHE) for synthetic data generation, enhancing local contrast while preserving image integrity levels. This method is expected to improve contrast in augmented images by 30%, simultaneously enhancing image quality metrics. Furthermore, the introduction of an overlapping sliding window technique with adaptive patch sizes ensures meticulous coverage and analysis of fundus images, significantly elevating lesion detection sensitivity and reducing false positives. Lastly, the framework employs an innovative multi-disease classification strategy utilizing ensemble learning and stacking of CNN architectures. This synergistic approach amalgamates multiple base models to diminish overfitting risks and enhance generalization capabilities, aiming for a classification accuracy surpassing 95% in binary assessments and 85% for specific retinal diseases. The proposed model not only addresses the current gaps in retinal image analysis but also sets a new standard for precision, reliability, and efficiency in diagnosing and managing ocular diseases, marking a significant step forward in the application of artificial intelligence in ophthalmology.

3D Convolutional Neural Networks for Predicting Protein Structure for Improved Drug Recommendation

INTRODUCTION: Protein structure prediction is critical for recommendation personalized medicine and drug discovery. This paper introduces a robust approach using 3D Convolution Neural Networks (3D CNN’s) to improve the accuracy of the structure of protein structure thus contributing for the drug recommendation system.
 OBJECTIVES: In contrast to conventional techniques, 3D CNNs are able to identify complicated folding patterns and comprehend the subtle interactions between amino acids because they are able to capture spatial dependencies inside protein structures.
 METHODS: Data sets are collected from Protein Data Bank, including experimental protein structures and the drugs that interact with them, are used to train the model. With the efficient processing of three-dimensional data, the 3D CNNs exhibit enhanced capability in identifying minute structural details that are crucial for drug binding. This drug recommendation system novel method makes it easier to find potential drugs that interact well with particular protein structures.
 RESULTS: The performance of the proposed classifier is compared with the existing baseline methods with various parameters accuracy, precision, recall, F1 score, mean squared error (MSE)  and area under the receiver operating characteristic curve (AUC-ROC).
 CONCLUSION: Deep learning and 3D structural insights work together to create a new generation of tailored and focused therapeutic interventions by speeding up the drug development process and improving the accuracy of pharmacological recommendations.

Coherent surface acoustic wave–spin wave interactions detected by micro-focused Brillouin light scattering spectroscopy

We investigated the interaction of surface acoustic waves and spin waves with spatial resolution by micro-focused Brillouin light scattering spectroscopy in a Co40Fe40B20(10 nm) ferromagnetic layer on a LiNbO3-piezoelectric substrate. We experimentally demonstrate that the magnetoelastic excitation of magnons by phonons is coherent by studying the interference of light scattered off generated magnons and annihilated phonons. We find a pronounced spatial dependence of the phonon annihilation and magnon excitation, which we map as a function of the magnetic field. The coupling efficiency of the surface acoustic waves (SAWs) and the spin waves is characterized by a magnetic field-dependent decay of the SAWs amplitude.

Micro-Doppler based Human Activity Recognition using ABOA based Dual Spatial Convolution with Gated Recurrent Unit

The through-wall capability, device-free detection of radar-based human activity recognition are drawing a lot of interest from both academics and industry. The majority of radar-based systems do not yet combine signal analysis and feature extraction in the frequency domain and the time domain. Applications like smart homes, assisted living, and monitoring rely on human identification and activity recognition (HIAR). Radar has a number of advantages over other sensing modalities, such as the ability to shield users' privacy and conduct contactless sensing. The article introduces a new human tracking system that uses radar and a classifier called Dual Spatial Convolution Gated Recurrent Unit (DSC-GRU) to identify the subject and their behavior. The system follows the person and identifies the type of motion whenever it detects movement. One important feature is the integration of the GRU with the DSC unit, which allows the model to simultaneously capture the spatiotemporal dependence. Present prediction models just take into account spatial features that are immediately adjacent to each other, disregarding or just superimposing global spatial features when taking spatial correlation into account. A new dependency graph is created by calculating the correlation among nodes using the correlation coefficient; this graph represents the global spatial dependence, while the classic static graph represents the neighboring spatial dependence in the DSC unit. The DSC unit goes a step further by using a modified gated mechanism to quantify the various contributions of both local and global spatial correlation. While previous models performed worse, the suggested model outperformed them with an accuracy of 99.45 percent and a precision of 97.15 percent.

The Impact of Land-Use Mix on Technological Innovation: Evidence from a Grid-Cell-Level Analysis of Shanghai, China

While the benefits of land-use mix have been widely analyzed with regard to transportation, public health, and economic development, relatively little attention has been paid to empirically investigating the impact of land-use mix on technological innovation at the intra-urban level. Drawing upon a database of geo-coded patents that are used to reflect the capacity of technological innovation, this paper takes Shanghai as a case study and analyzes how the intra-urban distribution of technological innovation has been associated with land-use mix at the 1 km × 1 km grid cell level. Empirical results, which are robust when the grids are divided at the 2 km × 2 km level, show that the degree of land-use mix is positively associated with the number of patents for a given grid, suggesting that grids with a higher level of land-use mix are likely to have more patents, ceteris paribus. Moreover, the results demonstrate an inverted U-curve relationship between land-use mix and technological innovation, indicating that a too much higher level of land-use mix could lead to a smaller number of patents for a certain grid. In addition, the empirical results suggest the existence of spatial dependence in the effect of land-use mix on technological innovation.

Directional dependency of mechanical properties in gallium phosphide semiconducting material

The aims of this contibution is to investigate the directional dependency of mechanical properties of gallium phosphide (GaP) semiconduct-ing material using the experimental stiffness elastic constants Cij reported in the litterature. These characteristics are visible by analysis and visualization of elastic tensors. The spatial dependencies of the Young’s modulus, linear compressibility, shear modulus, and Poisson’s ratio in GaP binary compound were calculated and schematized in graphical representations. Data analysis shows that the maximum value Emax of Young's modulus was estimated at around 166.7 GPa, while its minimum value Emin was estimated at around 102.5 GPa, which gives a ratio of approximately Emax/Emin ≈ 1.63, while for the Poisson’s ratio υ, the maximum value υmax of υ was estimated at around 0.43, while its minimum value υmin was estimated at 0.025, which gives a rate of approximately υmax/υmin ≈ 17.5. The Debye temperature was also calculated, it is estimated at about 482.4 K, it is in good agreement with the experimental one (493 K) reported in the litterature, and the deviation between these two values is less than 2.2%.

State-of-Health Estimation for Sustainable Electric Vehicle Batteries Using Temporal-Enhanced Self-Attention Graph Neural Networks

Abstract Electric vehicles (EVs) have emerged as an environmentally friendly alternative to conventional fuel vehicles. Lithium-ion batteries are the major energy source for EVs, but they degrade under dynamic operating conditions. Accurate estimation of battery state of health is important for sustainability as it quantifies battery condition, influences reuse possibilities, and helps alleviate capacity degradation, which finally impacts battery lifespan and energy efficiency. In this paper, a self-attention graph neural network combined with long short-term memory (LSTM) is introduced by focusing on using temporal and spatial dependencies in battery data. The LSTM layer utilizes a sliding window to extract temporal dependencies in the battery health factors. Two different approaches to the graph construction layer are subsequently developed: health factor-based and window-based graphs. Each approach emphasizes the interconnections between individual health factors and exploits temporal features in a deeper way, respectively. The self-attention mechanism is used to compute the adjacent weight matrix, which measures the strength of interactions between nodes in the graph. The impact of the two graph structures on the model performance is discussed. The model accuracy and computational cost of the proposed model are compared with the individual LSTM and gated recurrent unit (GRU) models.

A transportation Revitalization index prediction model based on Spatial-Temporal attention mechanism

The global outbreak of COVID-19 has had a substantial impact on various sectors worldwide, including the economy, healthcare, entertainment, policy formulation, and international relations, with the transportation industry being particularly hard-hit. To curb the widespread transmission of the virus, many regions globally have implemented policies and measures to restrict transportation. These actions not only directly affect the transportation industry but also further impose a severe impact on the economy and societal development of various areas. In this context, the Transportation Revitalization Index (TRI) becomes particularly important. It can evaluate the degree of recovery of city traffic conditions after the pandemic, and accurate prediction of TRI can help governments and decision-makers respond more precisely to the challenges that the pandemic brings to the transportation industry. However, existing research primarily focuses on the direct correlation between TRI change data and COVID-19 pandemic data, without fully considering the dynamic spatial correlation features and time dependency features that affect the nonlinear changes of TRI. In light of the above situation, this study proposes a Deep Spatial-Temporal prediction model based on the Attention Mechanism (DeepST-AM). The DeepST-AM deeply integrates historical TRI data with multivariate pandemic information and uses a spatial–temporal attention mechanism to capture the deep and complex spatial–temporal information of urban data. To more accurately capture the long-term complex features of TRI data, this paper designs a Gaussian temporal convolution model dedicated to TRI data. To validate the effectiveness of DeepST-AM, researchers used real data from 29 core cities in China as samples and compared the performance of DeepST-AM with existing multiple methods on TRI prediction tasks. The experimental results showed that compared to other methods, the DeepST-AM proposed in this paper has a significant advantage in the long-term prediction tasks of TRI in terms of performance evaluation, indicator prediction, etc. In summary, this research provides a more accurate and comprehensive prediction model for the traffic recovery status after the pandemic, hoping to provide strong support for future decisions.

River system changes under urban expansion: Insights from the Suzhou-Wuxi-Changzhou metropolitan area, China

Study regionThe Suzhou-Wuxi-Changzhou (SWC) metropolitan area, China. Study focusUrban expansion is a natural product of growing cities at the expense of river systems. In this study, we evaluated river system changes that come with urban expansion on a grid scale. We also utilized the spatial analysis models to examine the impact of urban expansion on river system changes. New hydrological insights for the regionOur results indicate that the cities in the metropolitan area had a similar pattern of urban expansion from the 1960s to the 2010s, with Changzhou, Wuxi, and Suzhou losing 45.8%, 32.3%, and 42.2% of their rivers, respectively. A negative correlation was detected between the urban expansion area (UEA) and the river density variation (∆Rd). During the rapid urban expansion period, ∆Rd significantly decreased as UEA increased. We then identified the spatial dependence and spatial spillover effect between river system changes and urban expansion. In our analysis, the river density was degraded as a result of urban expansion both in the in-situ location and in the neighboring locations. The river system has seen significant changes due to the impervious areas and infrastructures caused by urban expansion; thus, an appropriate pattern of urban development is required to mitigate the negative effects of river loss.

An estimation method for switching points of multimode spatiotemporal data based on SFSTAR

Various traditional methods, including statistical and spatial map-based approaches, are employed to model spatiotemporal data. However, the lack of explicit descriptions of spatial and temporal dependencies leads to unnecessarily intensive and uninterpretable relationships among the variables. To address this problem, space–time autoregressive moving average (STARMA)-type models that can depict temporal and spatial dependencies in a parametric manner are adopted in this study. However, such models may not be applicable to unlabeled multimode spatiotemporal data. In this study, we propose a switching flexible space–time autoregressive (SFSTAR) model to describe multimode spatiotemporal data. Considering the coupling between the parameter estimation of each mode and determination of the switching points, an estimation method is proposed to identify the mode switching points to partition the multimode data into different segments. First, the proposed estimation method identifies the number of segments by locating the transition regions that enclose the switching points based on the spatiotemporal feature extraction proposed herein. Second, with the estimated number of segments that is actually the number modes and the constraints on the locations of the switching points given by the transition regions, the proposed method jointly estimates the switching points and the SFSTAR model parameters in each mode through Bayesian optimization. The effectiveness of the proposed method is evaluated using a numerical example and a fiber optic temperature measurement dataset.

Multivariate time series anomaly detection with variational autoencoder and spatial–temporal graph network

Effective anomaly detection in multivariate time series (MTS) is very essential for modern complex physical equipment. A single anomaly in physical equipment can cause a series of failures due to fault propagation. Therefore, the equipment need to be comprehensively monitored by an anomaly detection system to ensure its health. However, given the rise in the size and complexity of the physical equipment, the number of sensors required to infer its health status has increased dramatically, and the complex temporal and spatial dependencies in sensors are difficult to capture. Aiming at the problem, this paper proposes a MTS anomaly detection algorithm called VSAD, which is based on spatial–temporal graph networks and variational autoencoder (VAE). It employs spatial–temporal graph networks to learn complex temporal and spatial dependencies in MTS. Moreover, its VAE architecture enables it to learn data distribution patterns in an unsupervised manner and its threshold are determined by an automatic threshold selection strategy. Finally, experimental results conducted on six publicly available benchmark datasets show that VSAD outperforms state-of-the-art methods for anomaly detection, with an average F1 scores improvement of about 7%.

Varying coefficient panel data models and methods under correlated error components: Application to disparities in mental health services in England

The contribution of this paper is twofold. Firstly, it introduces novel regression models that combine two important areas of the methodological development in panel data analysis, namely a varying coefficient specification and spatial error dependence. The former allows relatively flexible nonlinear interactions; the latter enables spatial correlations of the disturbance and thus differ significantly from the other random effect models in the literature. To estimate the model, a new estimation procedure is established that can be viewed as a generalization of the quasi-maximum likelihood method for a spatial panel data model to the well-known conditional local likelihood procedure. Novel inference methods, particularly variable selection and hypothesis testing of the parameter constancy, are introduced and are shown to be effective under the complex spatial error dependence. Equally importantly, this paper makes a substantial contribution to the understanding of financing and expenditure for health and social care. In particular, we empirically analyze and explain the effects of political ideologies on the local fiscal policy in England, especially the expenditure on mental health services.

Tracking the spatial footprints of extreme storm surges around the coastline of the UK and Ireland

Storm surges are the most important driver of flooding in many coastal areas. Understanding the spatial extent of storm surge events has important financial and practical implications for flood risk management, reinsurance, infrastructure reliability and emergency response. In this paper, we apply a new tracking algorithm to a high-resolution surge hindcast (CODEC, 1980–2017) to characterize the spatial dependence and temporal evolution of extreme surge events along the coastline of the UK and Ireland. We quantify the severity of each spatial event based on its footprint extremity to select and rank the collection of events. Several surge footprint types are obtained based on the most impacted coastal stretch from each particular event, and these are linked to the driving storm tracks. Using the collection of the extreme surge events, we assess the spatial distribution and interannual variability of the duration, size, severity, and type. We find that the northeast coastline is most impacted by the longest and largest storm surge events, while the English Channel experiences the shortest and smallest storm surge events. The interannual variability indicates that the winter seasons of 1989-90 and 2013–14 were the most serious in terms of the number of events and their severity, based on the return period along the affected coastlines. The most extreme surge event and the highest number of events occurred in the winter season 1989–90, while the proportion of events with larger severities was higher during the winter season 2013–14. This new spatial analysis approach of surge extremes allows us to distinguish several categories of spatial footprints of events around the UK/Ireland coast and link these to distinct storm tracks. The spatial dependence structures detected can improve multivariate statistical methods which are crucial inputs to coastal flooding assessments.

Spatial Functional Data analysis: Irregular spacing and Bernstein polynomials

Spatial Functional Data (SFD) analysis is an emerging statistical framework that combines Functional Data Analysis (FDA) and spatial dependency modeling. Unlike traditional statistical methods, which treat data as scalar values or vectors, SFD considers data as continuous functions, allowing for a more comprehensive understanding of their behavior and variability. This approach is well-suited for analyzing data collected over time, space, or any other continuous domain. SFD has found applications in various fields, including economics, finance, medicine, environmental science, and engineering. This study proposes new functional Gaussian models incorporating spatial dependence structures, focusing on irregularly spaced data and reflecting spatially correlated curves. The model is based on Bernstein polynomial (BP) basis functions and utilizes a Bayesian approach for estimating unknown quantities and parameters. The paper explores the advantages and limitations of the BP model in capturing complex shapes and patterns while ensuring numerical stability. The main contributions of this work include the development of an innovative model designed for SFD using BP, the presence of a random effect to address associations between irregularly spaced observations, and a comprehensive simulation study to evaluate models’ performance under various scenarios. The work also presents one real application of Temperature in Mexico City, showcasing practical illustrations of the proposed model.

A new approach for measuring poverty or social exclusion reduction in European NUTS 2 regions

This paper examines the problem of poverty or social exclusion reduction for NUTS 2 regions in the European Union over the years 2005-2015. To do so, we extend the concept of absolute poverty convergence to conditional poverty β-convergence and use a quantile regression approach to provide a broader assessment of the heterogeneous relationships between poverty and some economic factors. We consider spatial dependence between neighbouring regions using appropriate spatial econometric specification. We focus on regional performance in tackling poverty or social exclusion using a novel indicator based on the efficiency theory and examine potential heterogeneity in regions’ ability to fight poverty. These differences are analysed in detail considering instabilities in the relationship between regional performances and a relevant asset for policy effectiveness as the quality of institutions. Our findings can help policy makers to better understand factors behind poverty reductions and develop more responsive policies at different institutional levels.

Something new under the sun. A spatial econometric analysis of the adoption of photovoltaic systems in Italy

This paper analyses factors that determine the adoption of photovoltaic systems (PV) in Italy. We employ a spatial econometric approach applied to province-level data for 2015–2021 to identify the main determinants of PV adoption and to gauge both the potential bias deriving from spatial dependence and the spillover effects affecting neighboring areas in PV diffusion. We test different spatial econometric models and different types of economic, social and demographic variables. We add new regressors to consider the role of housing market dynamics (volume of sales and price of transaction) as a possible driver of PV adoption. We find that the housing market and electricity consumption are important positive determinants of PV adoption, whereas social capital and socio-demographic indicators do not provide statistical evidence of being major drivers of PV adoption. We confirm that other economic factors, like average income, can explain PV diffusion, whereas solar irradiation does not show to play a critical role. We do not find an important role of policies on energy efficiency and renewable energy in the housing sector, but our data timeframe covers a period after the very high incentives from the ‘Conto Energia’ and, at the same time, prevents us from observing the possible effects of the ‘Superbonus 110’ introduced in 2021. Moreover, our findings indicate that spatial dependence exists between neighboring areas in PV adoption, suggesting that spatial econometric models can be robust empirical approaches for interpreting PV deployment in studies at the regional or sub-regional level.

Parallel framework of a multi-graph convolutional network and gated recurrent unit for spatial–temporal metro passenger flow prediction

Metro passenger flow prediction is a critical problem in metro transport systems. However, recent studies have either overlooked spatial information on the metro network or primarily focused on modeling spatial dependencies using only the physical topology. To achieve accurate metro passenger flow (inflow and outflow at each station of a network) prediction, this study proposes a joint prediction model that combines the multi-graph convolution network and the gated recurrent unit (GRU). In addition to exploring location topology relationships, the proposed model selects two non-Euclidean spatial dependencies in metro passenger flow prediction to design essential graph elements as part of the stacked spatial block. Three spatial relationships (adjacency, similarity, and correlation) are integrated in parallel with the GRU network. The metro passenger flow prediction framework ASC-GRU (adjacency, similarity, correlation, and gated recurrent unit) is designed to mitigate the distortion of results during the capturing of passenger flow spatial–temporal features. Finally, ASC-GRU is tested using two datasets from the Hangzhou and Shanghai metro networks in China, and the error metrics of different models are compared and analyzed to verify the effectiveness and feasibility of ASC-GRU. The test results demonstrate that the proposed model outperforms other baseline models in passenger flow prediction over long time intervals and large networks. In particular, compared with the best performance of the baselines, the average reduction is around 3%, 12% and 13% in metrics of MAPE, MAE and RMSE, respectively.

Experimental investigation of the characteristics of the plasma flow generated in quasi-stationary plasma accelerator using optical methods

The spatial and temporal dependencies of the characteristics of the hydrogen plasma flow generated in quasi-stationary plasma accelerator were investigated. The spatiotemporal structure of discharge radiation in the interelectrode gap was studied. The range of changes in the length of the plasma glowing region in the interelectrode gap during the discharge pulse was determined. The region with bright plasma radiation located in the output face of the accelerator electrode system was observed. The presence of impurities and increased electron concentration values were observed in this region. Fluctuations in the radiation intensity of the plasma flow were detected along the entire length of its propagation. The spatial and temporal characteristics of these fluctuations were determined. The electron concentration values near the output face of the electrode system were obtained by measuring the Stark broadening of the Hβ line. For the first time, the time dependence of the electron concentration of free plasma flow was obtained using two methods simultaneously. The measurements were conducted at a distance, which significantly exceeds the characteristic size of the electrode system and where the influence of interelectrode processes of plasma flow generation is reduced. The first is based on measuring the Stark broadening of Hβ. As a second method, heterodyne interferometry was used.

Improving spatial transferability of deep learning models for small-field crop yield prediction

Predicting crop yield using deep learning (DL) and remote sensing is a promising technique in agriculture. In smallholder agriculture (<2 ha), where 84% of the farms operate globally, it is crucial to build a model that can be useful across several fields (high spatial transferability). However, enhancing spatial model transferability in a small-scale setting faces significant challenges, including spatial autocorrelation, heterogeneity and scale dependence of spatial dynamics, as well as the need to address limited data points. This study aimed to test the hypothesis that spatial cross validation (SCV) is a more suitable model validation practice than random cross validation (RCV) to enhance model transferability for spatial prediction in a small-scale farming setting. We compared the performances of DL models that predict crop yield for several settings including three crop types and two DL architectures based on RCV with and without overlapping samples and SCV. Notably, we conducted model performance tests on external, equally sized fields instead of the field used for training. We used high resolution RGB imagery taken with a drone as input. Our results show that the models using SCV outperformed those using RCV when the models were tested on external fields (on average r = 0.37 for SCV, r = 0.18 for RCV with overlap and r = 0.07 without), even though the models using SCV showed a substantially lower performance for cross validation (CV) than those using RCV (r with SCV and RCV w/o overlap = 0.73 and 0.98/0.73, respectively). The results suggest that RCV leads to over-optimism by overfitting the spatial structure and remembering image-specific information (so called memorization). Our study offers the first empirical evidence in agriculture that SCV is preferable to RCV in small field settings for making DL models more transferable.

Spatial-Temporal Graph Deviation Neural Networks for Anomaly Detection in Time Series

Abstract Anomaly detection in time series data (e.g., sensor data) is becoming a fundamental research problem that has various applications. Due to the complex inter-sensor relationships, it is challenging to detect anomalous events such as system faults and attacks hidden the high-dimensional time series. Recent advancements in deep learning approaches such as Graph Neural Networks (GNN) have greatly improved anomaly detection performance in time series data. However, existing methods usually use separate components to capture spatial and temporal dependence relationship and ignore the heterogeneities in spatial-temporal data which motivates us to capture them together. In this paper, we propose a novel approach STGDNN (short for Spatial-Temporal Graph Deviation Neural Networks) that learns the spatial-temporal dependence relationship together in structure learning of graph neural networks. In addition, we use the learned spatial-temporal graph structure and attention weights to explain the detected anomalies. The experiments on three real world datasets show our superiority in detection accuracy, anomaly diagnosis, and model interpretation compared with state-of-the-art methods.