RADSS: an Integration of GIS, Spatial Statistics, and Network Service for Regional Data Mining
Haitang Hu,Shuming Bao,Hui Lin,Qing Zhu
DOI: https://doi.org/10.1117/12.650384
2005-01-01
Abstract:Regional data mining, which aims at the discovery of knowledge about spatial patterns, clusters or association between regions, has widely applications nowadays in social science, such as sociology, economics, epidemiology, crime, and so on. Many applications in the regional or other social sciences are more concerned with the spatial relationship, rather than the precise geographical location. Based on the spatial continuity rule derived from Tobler's first law of geography: observations at two sites tend to be more similar to each other if the sites are close together than if far apart, spatial statistics, as an important means for spatial data mining, allow the users to extract the interesting and useful information like spatial pattern, spatial structure, spatial association, spatial outlier and spatial interaction, from the vast amount of spatial data or non-spatial data. Therefore, by integrating with the spatial statistical methods, the geographical information systems will become more powerful in gaining further insights into the nature of spatial structure of regional system, and help the researchers to be more careful when selecting appropriate models. However, the lack of such tools holds back the application of spatial data analysis techniques and development of new methods and models (e.g., spatio-temporal models). Herein, we make an attempt to develop such an integrated software and apply it into the complex system analysis for the Poyang Lake Basin. This paper presents a framework for integrating GIS, spatial statistics and network service in regional data mining, as well as their implementation. After discussing the spatial statistics methods involved in regional complex system analysis, we introduce RADSS (Regional Analysis and Decision Support System), our new regional data mining tool, by integrating GIS, spatial statistics and network service. RADSS includes the functions of spatial data visualization, exploratory spatial data analysis, and spatial statistics. The tool also includes some fundamental spatial and non-spatial database in regional population and environment, which can be updated by external database via CD or network. Utilizing this data mining and exploratory analytical tool, the users can easily and quickly analyse the huge mount of the interrelated regional data, and better understand the spatial patterns and trends of the regional development, so as to make a credible and scientific decision. Moreover, it can be used as an educational tool for spatial data analysis and environmental studies. In this paper, we also present a case study on Poyang Lake Basin as an application of the tool and spatial data mining in complex environmental studies. At last, several concluding remarks are discussed.