Maximum Getis–Ord Statistic Adjusted for Spatially Autocorrelated Data

Local statistics test the null hypothesis of no spatial association or clustering around the vicinity of a location. To carry out statistical tests, it is assumed that the observations are independent and that they exhibit no global spatial autocorrelation. In this article, approaches to account for global spatial autocorrelation are described and illustrated for the case of the Getis–Ord statistic with binary weights. Although the majority of current applications of local statistics assume that the spatial scale of the local spatial association (as specified via weights) is known, it is more often the case that it is unknown. The approaches described here cover the cases of testing local statistics for the cases of both known and unknown weights, and they are based upon methods that have been used with aspatial data, where the objective is to find changepoints in temporal data. After a review of the Getis–Ord statistic, the article provides a review of its extension to the case where the objective is to choose the best set of binary weights to estimate the spatial scale of the local association and assess statistical significance. Modified approaches that account for spatially autocorrelated data are then introduced and discussed. Finally, the method is illustrated using data on leukemia in central New York, and some concluding comments are made.     

Optimality in Two Spatially Separated Periodic Markets

A recursive analysis is made of a two market, trader-coupled, periodic marketing system with collapsing and recovering, nonlinear, demand and supply schedules. It yields time series in all prices and commodity flows. Conditions are established for maximum weekly social surplus, for the existence, uniqueness, and asymptotic stability of steady states, and for an optimal number of days for the marketing week.     

Spatial Distribution of Trees and Landscapes of the Past: A Mixed Spatially Correlated Multinomial Logit Model Approach for the Analysis of the Public Land Survey Data

Public Land Survey (PLS) data have been widely used in landscape studies of forest and woodlands in the pre‐ and early‐European‐settled Midwestern and Western United States. We aim to reconstruct presettlement forest vegetation at a finer spatial resolution than available from the PLS data using environmental covariates (slope, aspect, geology, and soil type) and the spatially correlated structure of witness tree data. To accommodate various data obtained from multiple sources while explicitly taking into account their spatial structures, we adopt a mixed spatially correlated multinomial logit model within the framework of a generalized linear mixed model. The application of the proposed model is illustrated using the three most abundant tree taxa from PLS data in the Arbuckle Mountains of south‐central Oklahoma. To assess the influence of each source of information on the spatial prediction, we considered four variant multinomial/spatial models and evaluated their relative predictive power using a validation technique. The probabilistic information about the spatial distribution of tree species obtained from different models reveals the need to integrate information about witness tree data as well as environmental covariates, and the nature of tree species; that is, a tendency to cluster in space to share environmental conditions in the reconstruction of the presettlement forest vegetation surface. Los datos sobre el uso y cobertura de tierras del Public Land Survey (PLS) han sido utilizados ampliamente en estudios de paisaje de bosques y de bosques históricos para periodo previo al asentamiento de migrantes europeos en el medio oeste y oeste de los Estados Unidos. Nuestro objetivo es reconstruir la vegetación forestal previa al asentamiento europeo a una resolución espacial más fina que la disponible actualmente en base a datos del PLS, usando covariables ambientales (pendiente, orientación, geología y tipo de suelo) y la estructura de correlación espacial de los datos de los árboles testigos. Para dar cabida a los diversos datos obtenidos de fuentes múltiples, y a la vez teniendo en cuenta explícitamente sus estructuras espaciales, adoptamos un modelo logit multinomial espacial mixto dentro del marco de los modelos mixtos lineales generalizados (GLMM). La aplicación del modelo propuesto es ilustrada con los tres tipos más abundantes de árboles según los datos del PLS para las montañas de Arbuckle en el centro‐sur de Oklahoma, EEUU. Para evaluar la influencia de cada fuente de información sobre la predicción espacial, se consideraron cuatro variantes de los modelos multinomial y espaciales. El poder predictivo de dichos modelos fue evaluado en relación con una técnica de validación. La información probabilística acerca de la distribución espacial de las especies de árboles obtenidos a partir de los diferentes modelos revela que para la reconstrucción de la superficie de la vegetación forestal histórica, es necesario integrar la información sobre los datos de árboles testigos así como las covariables ambientales y la naturaleza de las especies de árboles: es decir, la tendencia de los arboles a agruparse en el espacio para compartir las mismas condiciones ambientales. 公共土地调查(PLS)数据在欧洲人定居美国中西部和西部地区之前以及早期的森林和林地景观研究中得到广泛应用。本文旨在利用环境协变量（坡度、坡向、地貌和土地类型）证据树数据的空间关联结构,重建比PLS数据中更有效的更精细空间分辨率的前殖民期森林植被。为集成多种来源的各类数据,并明确地考虑数据间的空间结构,本文在广义线性混合模型(GLMM)框架下提出了混合空间关联多项Logit模型。以俄克拉荷马州中南部的阿尔布克尔山脉为研究区,提取PLS数据中三种最丰富的树种对模型进行验证。为估计每种信息来源对模型空间预测准确性的影响,本文考虑了4种变异的多项/空间模型并运用验证技术评估它们的相对预测能力。从不同模型获得的树种空间分布的概率信息表明,需要对证据树数据、环境协变量和树种自然属性信息进行集成,也就是说,在重建前殖民期森林植被曲面时,空间上的集聚趋势共享了环境条件。     

Location Equilibrium for Cournot Oligopoly in Spatially Separated Markets

Consider a two-stage non-cooperative Cournot game with location choice involving n≥ 2 firms each with several facilities. There are m≥ 2 spatially separated markets constituting the vertices of a network. Each firm first selects the locations of their facilities and then selects the quantities to supply to the markets to maximize its profit. There exists a Nash equilibrium in the quantities offered by each firm at the markets. Furthermore, when the demand in each market is sufficiently large, each firm chooses to locate its facilities only at vertices. With linear demand in each market, there exists a Nash location equilibrium.     

Simulation of Spatially Variable Seismic Underground Motions in U-Shaped Canyons

This paper focuses on the study of simulations for spatially variable seismic underground motions in U-shaped canyons. First, a canyon ground cross-coherence function based on commonly used coherence function models of flat terrain, is deduced and presented. To further obtain the underground cross-coherence function, a mathematical expression, including its specific deduction process for describing the relationship between coherence functions of multi-support ground and underground motions, is also given in detail and adopted. Then, the key factors (i.e. canyon underground power spectrum density and canyon underground coherence function) for simulating the spatially variable seismic underground motions are obtained by a two-step transferring method from flat-ground to underground soil. Furthermore, a program is coded for generating the spatially variable seismic underground motions. The effectiveness of the generated seismic motions is further verified. Finally, two numerical examples are taken to validate the proposed simulation approach, illustrating the specific characteristics of canyon coherence function. The analysis results show the apparent differences of the simulated seismic motions using the canyon coherence function from those using conventional coherence function models of flat terrain. The proposed approach provides some insights for anti-earthquake analysis of soil-structure interaction or underground structures in canyon topography.     

A Comparison of Spatially Varying Regression Coefficient Estimates Using Geographically Weighted and Spatial‐Filter‐Based Techniques

Geographically weighted regression (GWR) is a technique that explores spatial nonstationarity in data‐generating processes by allowing regression coefficients to vary spatially. It is a widely applied technique across domains because it is intuitive and conforms to the well‐understood framework of regression. An alternative method to GWR that has been suggested is spatial filtering, which it has been argued provides a superior alternative to GWR by producing spatially varying regression coefficients that are not correlated with each other and which display less spatial autocorrelation. It is, therefore, worthwhile to examine these claims by comparing the output from both methods. We do this by using simulated data that represent two sets of spatially varying processes and examining how well both techniques replicate the known local parameter values. The article finds no support that spatial filtering produces local parameter estimates with superior properties. The results indicate that the original spatial filtering specification is prone to overfitting and is generally inferior to GWR, while an alternative specification that minimizes the mean square error (MSE) of coefficient estimates produces results that are similar to GWR. However, since we generally do not know the true coefficients, the MSE minimizing specification is impractical for applied research.     

Comparison of the Spectral Representation Method to Simulate Spatially Variable Ground Motions

The spectral representation method (SRM) is widely used when simulating spatially variable ground motions. It has mainly two formulas, i.e., the random amplitudes and the random phases formulas. There exist three methods for decomposing the cross spectral density matrix: Cholesky decomposition, eigen decomposition, and root decomposition. Therefore, there are six forms with respect to the different combinations of the simulation formulas and the decomposition methods. To provide researchers and engineers with the guidance on choosing simulation method, the six forms are systematically investigated from five aspects: the power intensity, response spectra, and stochastic error of auto/cross spectral density, Fourier spectra, and difference indexes for Fourier amplitudes and phases. Finally, we give the following advice: the characteristics of the ground motions simulated by the random amplitudes formula are independent of the decomposition method, while the characteristics of the ground motions simulated by random phases formula are dependent of the decomposition method. Furthermore, the root decomposition is strongly recommended when utilizing the random phases formula.     

A Threshold-Satisfying Competitive Location Model

In this paper we consider a location model based on the threshold concept. We find the best location such that the probability of revenue falling short of the threshold is minimized. This objective is appropriate when a firm will not survive if its revenue falls below a known threshold. A new store is to be located. Demand is not deterministic but rather has a statistical distribution. We seek the location at which the probability that the revenue (expressed as market share attracted by the new store) is below a given threshold is minimized. The model is formulated and solved, and computational results are given.     

A Model of Lottery Participation

With the spread of state–sponsored lotteries to all but 17 states, it is important to understand the circumstances and perspectives of those who contribute to lottery revenues. A multivariate model of percent of income spent on lottery tickets indicates that participation is a declining function of income and education, and that it is higher among black, male, and older respondents. In addition, participation is affected by the social context of the respondent and the respondent's attitudes regarding the lottery as enjoyable and an escape. Attitudes favorable to lottery play are concentrated disproportionately among less advantaged groups, particularly the least educated.     

Categorical Wombling: Detecting Regions of Significant Change in Spatially Located Categorical Variables

Wombling is a method for discovering boundaries in a collection of continuous variables observed at the same geographic localities. We extend this method to categorical data. A categorical wombling statistic C_i, which identifies areas of rapid change, is defined for every pair i = 1,…, n of adjacent localities, and is equal to the average number of category mismatches at i. We use both simulation and theory to consider the order statistics of C_i under null hypotheses of randomness, and of spatial autocorrelation for each variable, but independence between variables. Graph-theoretical statistics derived from C_i investigate whether areas of rapid change resemble boundaries. Computer simulation is used to study the distributions of these under the two null hypotheses. The methods are applied to linguistic data in three European areas. Other potential applications exist in biology, linguistics, anthropology, and other social sciences.     

A Unified Model for Dispersing Facilities

One of the important classes of facility dispersion problems involves the location of a number of facilities where the intent is to place them as far apart from each other as possible. Four basic forms of the p‐facility dispersion problem appear in the literature. Erkut and Neuman present a classification system for these four classic constructs. More recently, Curtin and Church expanded upon this framework by the introduction of “multiple types” of facilities, where the dispersion distances between specific types are weighted differently. This article explores another basic assumption found in all four classic models (including the multitype facility constructs of Curtin and Church): that dispersion is accounted for in terms of either distance to the closest facility or distances to all facilities (from a given facility), whether applied to a single type of facility or across a set of facility types. In reality, however, measuring dispersion in terms of whether neighboring facilities to a given facility are dispersed rather than whether all facilities are dispersed away from the given facility often makes more sense. To account for this intermediate measure of dispersion, we propose a construct called partial‐sum dispersion. We propose four “partial‐sum” dispersion problem forms and show that these are generalized forms of the classic set of four models codified by Erkut and Neuman. Further, we present a unifying model that is a generalized form of all four partial‐sum models as well as a generalized form of the original four classic model constructs. Finally, we present computational experience with the general model and conclude with a few examples and suggestions for future research. Una de las clases importantes dentro de los problemas de dispersión de instalaciones de servicios/infraestructura es el caso en el que la localización de un número de instalaciones debe cumplir la condición de maximizar la distancia entre cada par. La literatura especializada cita cuatro formas básicas del problema de dispersión llamados tipo p‐instalación (p‐facility) (Shier 1977; Luna y Chaudhry 1984; Kuby 1987; Erkut y Neuman, 1991). Erkut y Neuman (1991) presentan un sistema de clasificación para estas cuatro formas clásicas. Recientemente, Curtin e Iglesia (2006) ampliaron este marco metodológico al incorporar múltiples tipos de instalaciones, permitiendo que las distancias de dispersión entre diferentes tipos específicos de instalaciones sean ponderadas de manera diferente. El artículo presente explora otro supuesto básico que se encuentra en los cuatro modelos clásicos (y las modifcaciones para acomodar instalaciones multi‐tipo de Curtin e Iglesia): la dispersión es cuantificada en términos de la distancia entre una instalación dada y la instalación más cercana, o entre una instalación dada y la totalidad de las instalaciones. Este supuesto se mantiene si las distancias son aplicadas a un solo tipo de instalación o a múltiples tipos de instalaciones. Sin embargo, en realidad, tiene más sentido medir la dispersión en relación a las instalaciones vecinas, en vez de en relación a la totalidad las instalaciones. Para incorporar esta realidad a un nuevo tipo de medida intermedia de dispersión, se propone una medida llamada dispersión de suma parcial (partial‐sum dispersion). Proponemos cuatro tipos de problemas de dispersión de tipo parcial‐sum y demostramos que éstas son formas generalizadas de los cuatro modelos clásicos presentados por Erkut y Neuman (1991). Además, se presenta un modelo unificado que es una forma generalizada de los cuatro modelos tipo partial‐sum, así como una forma generalizada de las cuatro tipos en el modelo clásico. Por último, se presenta los resultados de pruebas computacionales usando el modelo general y se concluye con algunos ejemplos y sugerencias para investigaciones futuras. 设施分散问题中重要的一类是大量设施的布局,其意图是将它们在空间上尽可能离得更远。目前文献中主要讨论了4种基本形式(Shier 1977; Moon and Chaudhry 1984; Kuby 1987; Erkut and Neuman 1991)。Erkut and Neuman (1991)提出了这4种经典结构的一种分类系统。Curtin and Church (2006)引入设施“多种类型”对上述分类框架进行拓展,在特定类别之间的分散距离的权重存在不同。本文探索了在4种经典模型中所发现的另一种基本假设（包含Curtin and Church的多种类型设施结构）：无论是在单一类型设施或包括多种类型设施中,分散度在解释某一给定设施到最近设施的距离或到所有设施的距离方面都是合理的。然而,在现实中设施分散度度量方面,测量某一给定设施的邻近设施的分散度特征相比于测量给定设施的所有其他设施的散布特征通常更有意义。为解释这种分散度的中间度量,本文提出了一种称为“局部和整体”的结构,包括4种分散问题形式,它们是Erkut and Neuman 4种传统类型的广义形式。本文进而提出了一个统一模型,即所有 “局部和整体”模型和经典类型结构一种广义形式。最后,对统一模型进行了计算检验,并基于几个实证进行了总结,还提出了未来的研究建议。