Yongwan Chun

Modeling network autocorrelation within migration flows by eigenvector spatial filtering

AbstractAlthough the assumption of independence among interaction flows frequently is engaged in spatial interaction modeling, in many circumstances it leads to misspecified models and incorrect inferences. An informed approach is to explicitly incorporate an assumed relationship structure among the interaction flows, and to explicitly model the network autocorrelation. This paper illustrates such an approach in the context of U.S. interstate migration flows. Behavioral assumptions, similar to those of the intervening opportunities or the competing destinations concepts, exemplify how to specify network flows that are related to particular origin–destination combinations. The stepwise incorporation of eigenvectors, which are extracted from a network link matrix, captures the network autocorrelation in a Poisson regression model specification context. Spatial autocorrelation in Poisson regression is measured by the test statistic of Jacqmin-Gadda et al. (Stat Med 16(11):1283–1297, 1997). Results show that estimated regression parameters in the spatial filtering interaction model become more intuitively interpretable.

Modeling Network Autocorrelation in Space–Time Migration Flow Data: An Eigenvector Spatial Filtering Approach

Gravity-type spatial interaction models have been popularly utilized in modeling cross-sectional migration data, but their misspecification also has been raised in the literature. This misspecification issue principally concerns an insufficient accounting of underlying effects of spatial structure, including the presence of network autocorrelation among migration flows. Recent studies reveal that spatial interaction models are significantly improved by incorporating network autocorrelation in log-linear or Poisson regression estimation techniques, which are common estimation methods for spatial interaction models. However, when migration flows are structured as a panel data set from multiple time periods, the data set is likely to display temporal correlation within each measurement unit (here, each flow between a dyad of an origin and a destination) as well as network autocorrelation within each time period. Hence, spatial interaction models should be explicitly specified to account for these two different types of correlation structure. Using the eigenvector spatial filtering technique, this article outlines how to model network autocorrelation among migration flows structured through multiple time spans in either a linear or a generalized linear mixed model. An analysis of annual U.S. interstate migration data reported by the U.S. Internal Revenue Service shows that incorporation of two different types of autocorrelation leads to an improvement of model fitting and more intuitive parameter estimates.

Spatial Autoregressive Model for Population Estimation at the Census Block Level Using LIDAR-derived Building Volume Information

The collection of population by census is laborious, time consuming and expensive, and often only available at limited temporal and spatial scales. Remote sensing based population estimation has been employed as a viable alternative for providing population estimates based on indicators that make use of two-dimensional areal information of buildings or one-dimensional length information of roads The recent advancement of LIDAR remote sensing provides the opportunity to add the third dimension of height information into the modeling of population distribution. This study explores the use of building volumes derived from LIDAR as a population indicator. Our study shows the volume-based model consistently outperforms area and length-based models at the census block level. Additionally, the study examines the impact of spatial autocorrelation, the presence of which violates the independence assumption of the traditional OLS models. To address this problem, a spatial autoregressive model is employed to account for the spatial autocorrelation in the regression residuals. By incorporating the spatial pattern, the volume-based spatial error model achieves a goodness of fit (R2) of 85 percent, with a significant improvement in model performance and estimation accuracies in comparison with its OLS counterpart. The study confirms building volume as a more valuable indicator and estimator for block level population distribution, especially if an appropriate spatial autoregressive model is adopted.

Modeling interregional commodity flows with incorporating network autocorrelation in spatial interaction models: An application of the US interstate commodity flows

Abstract Spatial interaction models are frequently used to predict and explain interregional commodity flows. Studies suggest that the effects of spatial structure significantly influence spatial interaction models, often resulting in model misspecification. Competing destinations and intervening opportunities have been used to mitigate this issue. Some recent studies also show that the effects of spatial structure can be successfully modeled by incorporating network autocorrelation among flow data. The purpose of this paper is to investigate the existence of network autocorrelation among commodity origin–destination flow data and its effect on model estimation in spatial interaction models. This approach is demonstrated using commodity origin–destination flow data for 111 regions of the United States from the 2002 Commodity Flow Survey. The results empirically show how network autocorrelation affects modeling interregional flows and can be successfully captured in spatial autoregressive model specifications.

Eigenvector selection with stepwise regression techniques to construct eigenvector spatial filters

Because eigenvector spatial filtering (ESF) provides a relatively simple and successful method to account for spatial autocorrelation in regression, increasingly it has been adopted in various fields. Although ESF can be easily implemented with a stepwise procedure, such as traditional stepwise regression, its computational efficiency can be further improved. Two major computational components in ESF are extracting eigenvectors and identifying a subset of these eigenvectors. This paper focuses on how a subset of eigenvectors can be efficiently and effectively identified. A simulation experiment summarized in this paper shows that, with a well-prepared candidate eigenvector set, ESF can effectively account for spatial autocorrelation and achieve computational efficiency. This paper further proposes a nonlinear equation for constructing an ideal candidate eigenvector set based on the results of the simulation experiment.

Delimitation of Functional Regions Using a p-Regions Problem Approach

Various spatial data analyses have been used for the identification of functional regions. Functional regions are identified by grouping many areal units into fewer clusters to classify the areal units in terms of similar properties, as well as to constrain the spatial contiguity of the areal units in each cluster. This article proposes a spatial optimization model, called the p-functional regions problem, to solve a regionalization problem by considering geographic flows. The magnitude of geographic flows, such as journey-to-work, is widely considered a good indicator of functional relationships between areas so that regionalization models incorporating various criteria, such as the maximum intraregion flows or the total inflows from other units, may be used to identify the p regions. We also propose an analytical target reduction approach to enhance the model tractability in generating optimal solutions to large problems and to demonstrate the effectiveness of the optimization model using journey-to-work data from Seoul (South Korea) and South Carolina (the United States).

The effects of different classification models on error propagation in land cover change detection

The use of land cover change maps is subject to the propagation of errors involved in classifying multi-temporal land cover maps. Understanding the link between classification processes and error propagation helps to determine appropriate classification models to mitigate the error propagation rate. In this paper, we present a simulation analysis on error propagation in land cover change detection using three classification models: a non-contextual model, a contextual model based on spatial smoothing, and a contextual model based on Markov random fields (MRF). A spatial simulation approach based on simulated annealing was developed with careful experimental designs to control two related factors including the spatial/temporal patterns of estimation errors associated with spectral probabilities. The results showed that the contextual classification model based on MRF had the smallest error propagation rate while the non-contextual classification model had the largest rate under all scenarios. The two factors had different effects on the error propagation for different classification models. For the non-contextual model, increasing temporal correlation of errors could reduce the error propagation rate while spatial autocorrelation of errors did not have a big impact on the error propagation. For the two contextual classification models, the use of contextual information significantly reduced the error propagation rate. However, the value of contextual information in mitigating error propagation was highly dependent on the spatial autocorrelation of the errors. The impact of the temporal correlation of errors was weakened in the contextual models.

Visualizing Migration Flows Using Kriskograms

This paper describes a new approach called kriskogram to visualizing migration flows. To create a kriskogram, geographical units are projected as a set of points on a straight line segment called a location line. The migration flow between two points on the location line is represented using a half-circle drawn from the origin to the destination in a clockwise direction. Translucent symbols and a classification scheme can be used to make a kriskogram more effective. We demonstrate this method using a set of interstate migration data of four time periods for the conterminous United States.

Network Reliability and Resilience of Rapid Transit Systems

The recent increase in demand and transportation security highlights the importance of the public transit system in the United States. This study explores how potential failures on nodal disruptions affect transit system flows and examines the change in the reliability of transit systems with a case study of the Greater Metropolitan Area of Washington, DC. For methodology, we employ network reliability and system flow loss and assess the criticality of stations under a variety of simulated nodal disruptions. We evaluate network resilience by identifying the best and worst geographical impact scenarios on networks.

Spatial Autocorrelation and Uncertainty Associated with Remotely-Sensed Data

Virtually all remotely sensed data contain spatial autocorrelation, which impacts upon their statistical features of uncertainty through variance inflation, and the compounding of duplicate information. Estimating the nature and degree of this spatial autocorrelation, which is usually positive and very strong, has been hindered by computational intensity associated with the massive number of pixels in realistically-sized remotely-sensed images, a situation that more recently has changed. Recent advances in spatial statistical estimation theory support the extraction of information and the distilling of knowledge from remotely-sensed images in a way that accounts for latent spatial autocorrelation. This paper summarizes an effective methodological approach to achieve this end, illustrating results with a 2002 remotely sensed-image of the Florida Everglades, and simulation experiments. Specifically, uncertainty of spatial autocorrelation parameter in a spatial autoregressive model is modeled with a beta-beta mixture approach and is further investigated with three different sampling strategies: coterminous sampling, random sub-region sampling, and increasing domain sub-regions. The results suggest that uncertainty associated with remotely-sensed data should be cast in consideration of spatial autocorrelation. It emphasizes that one remaining challenge is to better quantify the spatial variability of spatial autocorrelation estimates across geographic landscapes.