604
Views
14
CrossRef citations to date
0
Altmetric
Methods, Models, and GIS

Optimal Map Classification Incorporating Uncertainty Information

, &
Pages 575-590 | Received 01 May 2016, Accepted 01 Sep 2016, Published online: 13 Feb 2017
 

Abstract

A choropleth map frequently is used to portray the spatial pattern of attributes, and its mapping result heavily relies on map classification. Uncertainty in an attribute has an influence on map classification and, accordingly, can generate an unreliable spatial pattern. Only a few studies, however, have explored the implications of uncertainty in map classification. Recent studies present methods to incorporate uncertainty in map classification and generate a more reliable spatial pattern. Nevertheless, these methods often produce an undesirable result, with most observations assigned to one class, and struggle to find an optimal result. The purpose of this article is to expand the discussion about finding an optimal classification result considering data uncertainty in a map classification. Specifically, this article proposes optimal classification methods based on a shortest path problem in an acyclic network. These methods use dissimilarity measures and various cost and objective functions that simultaneously can consider attribute estimates and their uncertainty. Implementation of the proposed methods is in an ArcGIS environment with interactive graphic tools, illustrated with a mapping application of the American Community Survey data in Texas. The proposed methods successfully produce map classification results, achieving improved homogeneity within a class.

等值区域图经常用来描绘属性的空间模式,而其製图结果大幅倚赖地图的分级。属性中的不确定性,对地图分级具有影响,并且从而可能产生不可靠的空间模式。但仅有少数研究探讨地图分级中的不确定性之意涵。晚近的研究,呈现出将不确定性纳入地图分级的方法,并生产更为可靠的空间模式。但这些方法经常产生不良的结果,其中多半的观察被分配到单一级别,并难以找到最佳结果。本文的目的,便是扩张有关在考量地图分级中的数据不确定性之下寻找最适分级结果的讨论。本文特别提出根据非週期网络中的最短路径问题的最佳分级方法。这些方法运用不相似性测量与多种成本及目标函数,可同时考量属性判断及其不确定性。提出的方法,是在有互动图像工具的ArcGIS环境中实施,并以德州的美国社区调查数据的製图应用进行描绘。本文所提出的方法,成功地生产地图分级的结果,获得了改善的级别同质性。

Un mapa coroplético frecuentemente se usa para representar el patrón espacial de atributos, y su resultado cartográfico fuertemente se apoya en clasificación cartográfica. La incertidumbre en un atributo tiene una influencia sobre la clasificación cartográfica y, en consecuencia, puede llevar a la generación de un patrón espacial poco fiable. Sin embargo, solo unos pocos estudios han explorado las implicaciones de incertidumbre en clasificación cartográfica Estudios recientes presentan métodos para incorporar la incertidumbre en clasificación cartográfica y generar un patrón espacial más confiable. Aun así, estos métodos a menudo producen un resultado indeseable, con la mayoría de las observaciones asignadas a una clase, y bregan por encontrar un resultado óptimo. El propósito de este artículo es ampliar la discusión sobre cómo encontrar un resultado de clasificación óptimo considerando la incertidumbre de los datos en clasificación cartográfica. Específicamente, este artículo propone métodos de clasificación óptimos basados en un problema de la ruta más corta en una red en espiral. Estos métodos usan medidas de disimilitud y varias funciones de costo y objetivo que simultáneamente pueden considerar los estimativos de atributo y su incertidumbre. La implementación de los métodos propuestos se hace en un entorno ArcGIS con herramientas gráficas interactivas, ilustradas con una aplicación de mapeo de datos del Estudio Comunitario Americano en Texas. Los métodos propuestos con todo éxito producen resultados de clasificación cartográfica, logrando una homogeneidad mejorada dentro de una clase.

Funding

This research was supported by the National Institutes of Health, Grant 1R01HD076020-01A1. Any opinions, findings, and conclusions or recommendations expressed in this article are those of the authors and do not necessarily reflect the views of the National Institutes of Health.

Notes

1. The standard two-sample z test compares two means with sampling distributions that conform to a normal distribution with known standard deviations.

2. The range of observation counts is calculated as the difference between the largest and smallest numbers of observations in classes.

Additional information

Notes on contributors

Hyeongmo Koo

HYEONGMO KOO is a PhD student in the School of Economic, Political and Policy Sciences at the University of Texas at Dallas, Richardson, TX 75080. E-mail: [email protected]. His research interests include spatial data uncertainty and geovisualization.

Yongwan Chun

YONGWAN CHUN is an Associate Professor of Geospatial Information Sciences at the University of Texas at Dallas, Richardson, TX 75080. E-mail: [email protected]. His research interests lie in spatial statistics and GIS focusing on urban issues including population movement, environment, health, and crime.

Daniel A. Griffith

DANIEL A. GRIFFITH is Ashbel Smith Professor of Geospatial Information Sciences at the University of Texas at Dallas, Richardson, TX 75080. E-mail: [email protected]. He also is a former Guggenheim fellow and has been awarded distinguished scholarship honors by the American Association of Geographers. He has published nearly two dozen books and more than 200 papers and is a previous editor of Geographical Analysis.

Reprints and Corporate Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

To request a reprint or corporate permissions for this article, please click on the relevant link below:

Academic Permissions

Please note: Selecting permissions does not provide access to the full text of the article, please see our help page How do I view content?

Obtain permissions instantly via Rightslink by clicking on the button below:

If you are unable to obtain permissions via Rightslink, please complete and submit this Permissions form. For more information, please visit our Permissions help page.