Please enable JavaScript to view this site.
bg

Benny Chin 陳威全

a Geographer, Cartographer, & GIScientist

A Critical Scale for Analyzing Point Patterns: An Analysis of Dengue Fever Cases in Kaohsiung City

Wei Chien Benny Chin  (2022)

Paper

Wei Chien Benny Chin (2022). A Critical Scale for Analyzing Point Patterns: An Analysis of Dengue Fever Cases in Kaohsiung City, 人口學刊. 65: 1--41.

Abstract

In a disease outbreak context, disease cases are usually presented by using point distribution data. Due to the scale-invariant issue of point data and the scaling issue of the modifiable areal unit problem, identifying a critical scale for the analysis of point patterns, such as the clustering phenomenon, is important. This study proposes a novel data-driven framework for calculating the critical scale based on two traditional concepts: (1) the point-region quadtree spatial indexing method and (2) the box counting method for fractal pattern analysis. Both concepts capture the spatial scaling process and serve as the core concepts of the proposed framework. Using dengue fever cases in Kaohsiung City, Taiwan, during the past two decades, the critical scale was identified for each outbreak year. Two clustering analysis approaches were used to test the resulting critical scales, including kernel density estimation and density-based spatial clustering application with noise. Both clustering analyses involved distance parameter settings. Therefore, through the setting of search radii, the two clustering methods were used as a tool to explore the clustering patterns under different scale levels. In summary, the identified critical scales can better capture the spatial patterns of point data.

English info

Chin, W. C. B. (2022). The Critical Scale for Analyzing Point Patterns: An Analysis of Dengue Fever Cases in Kaohsiung City (in english, TSSCI). Journal of Population Studies 65: 1--41.

分析點分布的臨界尺度: 以高雄市登革熱病例的點資料為例

摘要

疾病的空間群聚分析中常用點資料型態對病例分布位置進行標記與分析。因為點資料的無尺度特性與可調整面單元問題,計算出適當的分析尺度對進行點分布的群聚分析非常關鍵。本研究提出一個資料導向的新分析架構以計算出臨界尺度。此架構是基於兩項經典分析概念:一、點區塊四分樹空間索引方法,二、碎型型態分析中的計算格子法。這兩者都具有捕捉空間跨尺度過程的特性,而這跨尺度特性即為此分析架構的核心概念。本研究應用臺灣高雄市過去近20年的登革熱病例資料作為分析對象,應用前述分析架構計算出每一波疫情年的臨界尺度。透過兩種群聚分析方法,包括核密度推估法與基於密度群聚與噪點分析法,本研究對臨界尺度所計算出來的群聚結構與其他尺度計算所得之群聚結構進行比較與討論。這兩項群聚分析都需設定一距離參數,即其搜尋半徑。因此,本研究透過設定不同的搜尋半徑,借用兩項群聚分析方法所得的群聚型態作為比較不同尺度下的點分布型態以進行比較討論。兩項群聚分析的比較結果都顯示在臨界尺度所計算的群聚結構特性相對較可突出其分布型態。

Read article here: