Multivariate Spatial Process Research Articles

LAND-USE FILTERING FOR NONSTATIONARY SPATIAL PREDICTION OF COLLECTIVE EFFICACY IN AN URBAN ENVIRONMENT.

Collective efficacy-the capacity of communities to exert social control toward the realization of their shared goals-is a foundational concept in the urban sociology and neighborhood effects literature. Traditionally, empirical studies of collective efficacy use large sample surveys to estimate collective efficacy of different neighborhoods within an urban setting. Such studies have demonstrated an association between collective efficacy and local variation in community violence, educational achievement, and health. Unlike traditional collective efficacy measurement strategies, the Adolescent Health and Development in Context (AHDC) Study implemented a new approach, obtaining spatially-referenced, place-based ratings of collective efficacy from a representative sample of individuals residing in Columbus, OH. In this paper we introduce a novel nonstationary spatial model for interpolation of the AHDC collective efficacy ratings across the study area, which leverages administrative data on land use. Our constructive model specification strategy involves dimension expansion of a latent spatial process and the use of a filter defined by the land-use partition of the study region to connect the latent multivariate spatial process to the observed ordinal ratings of collective efficacy. Careful consideration is given to the issues of parameter identifiability, computational efficiency of an MCMC algorithm for model fitting, and fine-scale spatial prediction of collective efficacy.

Statistical Depth in Spatial Point Process

Statistical depth is widely used as a powerful tool to measure the center-outward rank of multivariate and functional data. Recent studies have introduced the notion of depth to the temporal point process, which exhibits randomness in the cardinality as well as distribution in the observed events. The proposed methods can well capture the rank of a point process in a given time interval, where a critical step is to measure the rank by using inter-arrival events. In this paper, we propose to extend the depth concept to multivariate spatial point process. In this case, the observed process is in a multi-dimensional location and there are no conventional inter-arrival events in the temporal process. We adopt the newly developed depth in metric space by defining two different metrics, namely the penalized metric and the smoothing metric, to fully explore the depth in the spatial point process. The mathematical properties and the large sample theory, as well as depth-based hypothesis testings, are thoroughly discussed. We then use several simulations to illustrate the effectiveness of the proposed depth method. Finally, we apply the new method in a real-world dataset and obtain desirable ranking performance.

CoYangCZ: a new spatial interpolation method for nonstationary multivariate spatial processes

In multivariate spatial interpolation, the accuracy of a variable of interest can be improved using ancillary variables. Although geostatistical methods are widely used for multivariate spatial interpolation, these methods usually require second-order stationary assumption of spatial processes, which is difficult to satisfy in practice. We developed a new multivariate spatial interpolation method based on Yang-Chizhong filtering (CoYangCZ) to overcome this limitation. CoYangCZ does not solve the multivariate spatial interpolation problem from a purely statistical point of view but integrates geometry and statistics-based strategies. First, we used a weighted moving average method based on binomial coefficients (i.e. Yang-Chizhong filtering) to fit the spatial autocorrelation structure of each spatial variable from a geometric perspective. We then quantified the spatial autocorrelation of each spatial variable and the correlations between different spatial variables by analyzing the variances of different spatial variables. Finally, we obtain the best linear unbiased estimators at the unsampled locations. Experiments on air pollution and meteorological datasets show that CoYangCZ has a higher interpolation accuracy than cokriging, regression kriging, gradient plus-inverse distance squared, sequential Gaussian co-simulation, and the kriging convolutional network. CoYangCZ can adapt to second-order non-stationary spatial processes; therefore, it has a wider scope of application than purely statistical methods.

Modeling Massive Highly Multivariate Nonstationary Spatial Data with the Basis Graphical Lasso

We propose a new modeling framework for highly multivariate spatial processes that synthesizes ideas from recent multiscale and spectral approaches with graphical models. The basis graphical lasso writes a univariate Gaussian process as a linear combination of basis functions weighted with entries of a Gaussian graphical vector whose graph is estimated from optimizing an l 1 penalized likelihood. This article extends the setting to a multivariate Gaussian process where the basis functions are weighted with Gaussian graphical vectors. We motivate a model where the basis functions represent different levels of resolution and the graphical vectors for each level are assumed to be independent. Using an orthogonal basis grants linear complexity and memory usage in the number of spatial locations, the number of basis functions, and the number of realizations. An additional fusion penalty encourages a parsimonious conditional independence structure in the multilevel graphical model. We illustrate our method on a large climate ensemble from the National Center for Atmospheric Research’s Community Atmosphere Model that involves 40 spatial processes. Supplementary materials for this article are available online.

Asymptotic properties of nonparametric quantile estimation with spatial dependency

The purpose of this work is to nonparametrically estimate the conditional quantile for a locally stationary multivariate spatial process. The new kernel quantile estimate derived from the one of conditional distribution function (CDF). The originality in the paper is based on the ability to take into account some local spatial dependency in estimate CDF form. Consistency and asymptotic normality of the estimates are obtained under ‐mixing condition. Numerical study and application to real data are given in order to illustrate the performance of our methodology.

Quasi-likelihood for multivariate spatial point processes with semiparametric intensity functions

We propose a new estimation method to fit a semiparametric intensity function model to multivariate spatial point processes. Our approach is based on the so-called quasi-likelihood that can produce more efficient estimators by accounting for both between- and within-process correlations. To be more specific, we derive the optimal estimating function in a class of first-order estimating functions, where the optimal estimating function depends on the solution to a system of integral equations. We propose a computationally fast approach to obtain an approximate solution to the integral equation, and the resulting estimation approach is therefore computationally efficient. We demonstrate the efficacy of the proposed approach through both simulations and a real application.

Graphical Gaussian Process Models for Highly Multivariate Spatial Data.

For multivariate spatial Gaussian process (GP) models, customary specifications of cross-covariance functions do not exploit relational inter-variable graphs to ensure process-level conditional independence among the variables. This is undesirable, especially for highly multivariate settings, where popular cross-covariance functions such as the multivariate Matérn suffer from a "curse of dimensionality" as the number of parameters and floating point operations scale up in quadratic and cubic order, respectively, in the number of variables. We propose a class of multivariate "Graphical Gaussian Processes" using a general construction called "stitching" that crafts cross-covariance functions from graphs and ensures process-level conditional independence among variables. For the Matérn family of functions, stitching yields a multivariate GP whose univariate components are Matérn GPs, and conforms to process-level conditional independence as specified by the graphical model. For highly multivariate settings and decomposable graphical models, stitching offers massive computational gains and parameter dimension reduction. We demonstrate the utility of the graphical Matérn GP to jointly model highly multivariate spatial data using simulation examples and an application to air-pollution modelling.

The effective sample size for multivariate spatial processes with an application to soil contamination

AbstractEffective sample size accounts for the equivalent number of independent observations contained in a sample of correlated data. This notion has been widely studied in the context of univariate spatial variables. In that case, the effective sample size determines the reduction in the sample size due to the existing spatial correlation. In this paper, we generalize the methodology for multivariate spatial variables to provide a common effective sample size when all variables have been measured at the same locations. Together with the definition, we provide examples to investigate what an effective sample size looks like. An application for a soil contamination data set is considered. To reduce the dimensions of the process, clustering techniques are applied to obtain three bivariate vectors that are modeled using coregionalization models. Because the sample size of the data set is moderate and the locations are very unevenly distributed in the study area, the spatial analysis is challenging and interesting. We find that due to the presence of spatial autocorrelation, the sample size can be reduced by 38.53%, avoiding the duplication of information.Recommendations for Resource Managers: Before carrying out a sample survey with georeferenced data, it is essential to consider the impact of spatial correlation on sample size calculations. When the nature of the problem requires multivariate characteristics analysis, we provide a methodology to evaluate the effective sample size from a multivariate perspective. If the sample size is large, the effective sample size allows us to define the size of the subsample that should be used to preserve the theoretical properties of the estimation of the mean.

New Robust Cross-Variogram Estimators and Approximations of Their Distributions Based on Saddlepoint Techniques

Let Z(s)=(Z1(s),…,Zp(s))t be an isotropic second-order stationary multivariate spatial process. We measure the statistical association between the p random components of Z with the correlation coefficients and measure the spatial dependence with variograms. If two of the Z components are correlated, the spatial information provided by one of them can improve the information of the other. To capture this association, both within components of Z(s) and across s, we use a cross-variogram. Only two robust cross-variogram estimators have been proposed in the literature, both by Lark, and their sample distributions were not obtained. In this paper, we propose new robust cross-variogram estimators, following the location estimation method instead of the scale estimation one considered by Lark, thus extending the results obtained by García-Pérez to the multivariate case. We also obtain accurate approximations for their sample distributions using saddlepoint techniques and assuming a multivariate-scale contaminated normal model. The question of the independence of the transformed variables to avoid the usual dependence of spatial observations is also considered in the paper, linking it with the acceptance of linear variograms and cross-variograms.

Spatial factor modeling: A Bayesian matrix-normal approach for misaligned data.

Multivariate spatially oriented data sets are prevalent in the environmental and physical sciences. Scientists seek to jointly model multiple variables, each indexed by a spatial location, to capture any underlying spatial association for each variable and associations among the different dependent variables. Multivariate latent spatial process models have proved effective in driving statistical inference and rendering better predictive inference at arbitrary locations for the spatial process. High-dimensional multivariate spatial data, which are the theme of this article, refer to data sets where the number of spatial locations and the number of spatially dependent variables is very large. The field has witnessed substantial developments in scalable models for univariate spatial processes, but such methods for multivariate spatial processes, especially when the number of outcomes are moderately large, are limited in comparison. Here, we extend scalable modeling strategies for a single process to multivariate processes. We pursue Bayesian inference, which is attractive for full uncertainty quantification of the latent spatial process. Our approach exploits distribution theory for the matrix-normal distribution, which we use to construct scalable versions of a hierarchical linear model of coregionalization (LMC) and spatial factor models that deliver inference over a high-dimensional parameter space including the latent spatial process. We illustrate the computational and inferential benefits of our algorithms over competing methods using simulation studies and an analysis of a massive vegetation index dataset.

Bayesian Spatial Modeling for Housing Data in South Africa.

Spatial process models are being increasingly employed for analyzing data available at geocoded locations. In this article, we build a hierarchical framework with multivariate spatial processes, where the outcomes are "mixed" in the sense that some may be continuous, some binary and others may be counts. The underlying idea is to build a joint model by hierarchically building conditional distributions with different spatial processes embedded in each conditional distribution. The idea is simple and the resulting models can be fitted to multivariate spatial data using straightforward Bayesian computing methods such as Markov chain Monte Carlo methods. Bayesian inference is carried out for parameter estimation and spatial interpolation. The proposed models are illustrated using housing data collected in the Walmer district of Port Elizabeth, South Africa. Inferential interest resides in modeling spatial dependencies of dependent outcomes and associations accounting for independent explanatory variables. Comparisons across different models confirm that the selling price of a house in our data set is relatively better modeled by incorporating spatial processes.

Statistical challenges in spatial analysis of plant ecology data

The scope of spatial statistics problems arising in ecology is substantial. They concern both plant and animal behavior. Here, we discuss only plants. The range of issues is still very large and so we confine ourselves to four challenges, each within the context of model development. These are: species distribution models, both individual and joint; preferential sampling and bias in presence/absence data collection; time to event data across spatial regions; high-dimensional multivariate spatial process model fitting in the context of joint species distribution models.

Analysing Multivariate Spatial Point Processes with Continuous Marks: A Graphical Modelling Approach

SummaryThis paper discusses the application of graphical modelling in the statistical analysis of marked point patterns. We consider a multivariate planar point process with quantitative marks. After a survey of statistical methods for marked point processes, a new graphical model is presented. The sub‐processes of marked points with identical discrete marks are identified with nodes of a graph, which is used to describe aspects of the spatial relationship: If two sub‐patterns are similar, then an arc is made between the corresponding nodes. Similarity is defined based on spectral densities, which makes the computations efficient. The resulting graph presents all these pairwise similarities simultaneously. We demonstrate the application of our method in the analysis of a multi‐species forest, where the points are tree locations, the discrete marks tree species and the quantitative marks are the diameters at breast height.

Bayesian Spatial Multivariate Receptor Modeling for Multisite Multipollutant Data

ABSTRACTFor the development of effective air pollution control strategies, it is crucial to identify the sources that are the principal contributors to air pollution and estimate how much each source contributes. Multivariate receptor modeling aims to address these problems by decomposing ambient concentrations of multiple air pollutants into components associated with different source types. With the expanded monitoring efforts that have been established over the past several decades, extensive multivariate air pollution data obtained from multiple monitoring sites (multisite multipollutant data) are now available. Although considerable research has been conducted on modeling multivariate space-time data in other contexts, there has been little research on spatial multivariate receptor models for multisite, multipollutant data. We present a Bayesian spatial multivariate receptor modeling (BSMRM) approach that can incorporate spatial correlations in multisite, multipollutant data into the estimation of source composition profiles and contributions, based on discrete process convolution models for multivariate spatial processes. The new BSMRM approach enables predictions of source contributions at unmonitored sites as well as simultaneously dealing with model uncertainty caused by the unknown number of sources and identifiability conditions. The new approach can also provide uncertainty estimates for the predicted source contributions at any location, which was not possible in previous multivariate receptor modeling approaches. The proposed approach is applied to 24-hour ambient air concentrations of 17 Volatile Organic Compounds (VOCs) measured at nine monitoring sites in Harris County, Texas, between 2003 and 2005. Supplementary materials for this article, including real data and MATLAB codes for implementing BSMRM, are available online on the journal web site.

Multivariate spatial meta kriging

This work extends earlier work on spatial meta kriging for the analysis of large multivariate spatial datasets as commonly encountered in environmental and climate sciences. Spatial meta-kriging partitions the data into subsets, analyzes each subset using a Bayesian spatial process model and then obtains approximate posterior inference for the entire dataset by optimally combining the individual posterior distributions from each subset. Importantly, as is often desired in spatial analysis, spatial meta kriging offers posterior predictive inference at arbitrary locations for the outcome as well as the residual spatial surface after accounting for spatially oriented predictors. Our current work explores spatial meta kriging idea to enhance scalability of multivariate spatial Gaussian process model that uses linear model co-regionalization (LMC) to account for the correlation between multiple components. The approach is simple, intuitive and scales multivariate spatial process models to big data effortlessly. A simulation study reveals inferential and predictive accuracy offered by spatial meta kriging on multivariate observations.

Bayesian Spectral Modeling for Multivariate Spatial Distributions of Elemental Concentrations in Soil

Recent technological advances have enabled researchers in a variety of fields to collect accurately geocoded data for several variables simultaneously. In many cases it may be most appropriate to jointly model these multivariate spatial processes without constraints on their conditional relationships. When data have been collected on a regular lattice, the multivariate conditionally autoregressive (MCAR) models are a common choice. However, inference from these MCAR models relies heavily on the pre-specified neighborhood structure and often assumes a separable covariance structure. Here, we present a multivariate spatial model using a spectral analysis approach that enables inference on the conditional relationships between the variables that does not rely on a pre-specified neighborhood structure, is non-separable, and is computationally efficient. Covariance and cross-covariance functions are defined in the spectral domain to obtain computational efficiency. The resulting pseudo posterior inference on the correlation matrix allows for quantification of the conditional dependencies. A comparison is made with an MCAR model that is shown to be highly sensitive to the choice of neighborhood. The approaches are illustrated for the toxic element arsenic and four other soil elements whose relative concentrations were measured on a microscale spatial lattice. Understanding conditional relationships between arsenic and other soil elements provides insights for mitigating pervasive arsenic poisoning in drinking water in southern Asia and elsewhere.

Control charts for multivariate spatial autoregressive models

This paper deals with spatial detection of changes in model parameters of spatial autoregressive processes. The respective sequential testing problems are formulated. Moreover, we introduce characteristic quantities to monitor means or covariances of multivariate spatial autoregressive processes. Additionally, we also take into account the simultaneous surveillance of the mean vector and the covariance matrix. The aim is to apply control charts, important tools of sequential analysis, to these quantities. The considered control procedures are based on either cumulative sums or exponential smoothing. Further, we illustrate the methodology of statistical process control studying the spectrum of additive colors in a satellite photograph. Via simulation studies, the proposed control procedures are calibrated for a predefined average run length. In addition, we compare the performance of the control procedures considering the out-of-control situation. Eventually, the control charts are applied, and the signals of the different schemes are visualized. The final results are critically discussed.

On the smallest eigenvalues of covariance matrices of multivariate spatial processes

There has been a growing interest in providing models for multivariate spatial processes. A majority of these models specify a parametric matrix covariance function. Based on observations, the parameters are estimated by maximum likelihood or variants thereof. While the asymptotic properties of maximum likelihood estimators for univariate spatial processes have been analyzed in detail, maximum likelihood estimators for multivariate spatial processes have not received their deserved attention yet. In this article, we consider the classical increasing‐domain asymptotic setting restricting the minimum distance between the locations. Then, one of the main components to be studied from a theoretical point of view is the asymptotic positive definiteness of the underlying covariance matrix. Based on very weak assumptions on the matrix covariance function, we show that the smallest eigenvalue of the covariance matrix is asymptotically bounded away from zero. Several practical implications are discussed as well. Copyright © 2016 John Wiley & Sons, Ltd.

Modeling Brand Correlation Through Iterative Sparsity Search

The quantification of brand correlation is an important task in marketing research to understand the similarity of sales performance between multiple brands of similar products. We propose a statistical model for studying the sales performance of a single brand in multiple geographic locations (markets) as a spatial process, and the similarity between two brands is quantified as the covariance of two spatial processes. For fast computing that scales well to a large number of brands and markets, we propose a new algorithm called iterative sparsity search (ISS), to efficiently estimate a sparse covariance structure in the multivariate spatial model. We demonstrate the effectiveness of the proposed method using real-world data from consumer packaged goods in grocery stores. In this case study, the sales performance of multiple brands in multiple markets is modeled using multivariate spatial processes with a covariance structure incorporating both brand dependence and spatial dependence. The holdout-sample results in the study show that the brand correlation estimates from the proposed method are very useful in predicting the sales performance of grocery brands in unexplored markets.

A multivariate geostatistical approach for landscape classification from remotely sensed image data

This paper proposes a methodology to address the classification of images that have been acquired from remote sensors. One common problem in classification is the high dimensionality of multivariate characteristics. The methodology we propose consists of reducing the dimensionality of the spectral bands associated with a multispectral satellite image. Such dimensionality reduction is accomplished by the use of the divergence of a modified Mahalanobis distance. Instead of using the covariance matrix of a multivariate spatial process, the codispersion matrix is considered which have some desirable asymptotic properties under very precise conditions. The consistency and asymptotic normality hold for a general class of processes that are a natural extension of the one-dimensional spatial processes for which the asymptotic properties were first established. The results allow the selection of a set of spectral bands to produce the highest value of divergence. Then, a supervised maximum likelihood method using the selected spectral bands is employed for landscape classification. An application with a real LANDSAT image is introduced to explore and visualize how our method works in practice.