|
|
|
| COMPARATIVE ANALYSIS OF AGGREGATION DETECTION BASED ON SPATIAL AUTOCORRELATION AND SPATIAL-TEMPORAL SCAN STATISTICS |
| WANG Pei-an1, LUO Wei-hua2, BAI Yong-ping1 |
1. College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, China;
2. The 401st Institute of the Fourth Academy of CASC, Xi'an 710025, China |
|
|
|
|
Abstract The problem of aggregation economy has been always one of the hot and focuses of the regional economic research, and the positioning of aggregation is a prerequisite for continuing in-depth study in aggregation analysis. Using different methods, the scope of the aggregation phenomenon is different. Because aggregation is high sensitivity in scale, the influence of scale must be fully considered in using the spatial analysis methods. Using spatial autocorrelation methods to detect aggregation, the scale of choice is often susceptible to subjective judgments of the researchers; there is the issue of possible with selection bias, so the spatial weight problem has been controversial. For spatial autocorrelation analysis, the spatial weight matrix creation is the most difficult problems. Using different spatial weight matrix will get different results. In addition, the aggregation is obviously space-dependent as well as time-dependent since different aggregation takes place within different time and space, which is neglected by the spatial autocorrelation method. This is a concrete manifestation of the difference in time and space. So space-time analysis is necessarily a method which must be selected. At this point, the scan statistics method which was developed by Kulldorff and other scholars demonstrated a unique advantage. The exploratory research will use data about industrial population in some cities and counties of Zhejiang from 2000 to 2009. In particular, by comparing the space-time scan statistics method and a simple spatial autocorrelation method, which effectively introduces a time variable, this approach allows researchers to solve the problem a long time, but this approach can also increase the accuracy of the results. This shows that it is not only effective measure to solve the problem of artificial selection, to achieve the Scale extrapolation and automatic conversion, and more favorable mix of three-dimensional, dynamic, multi-scale analysis.
|
|
Received: 02 September 2011
|
|
|
|
|
|
|
2012, Vol. 27 
