摘要
应用模糊联系度的概念,以探讨对差异度系数进行数据挖掘的改进集对分析方法,实现对不确定性数据信息的数据挖掘和深入分析,提高评价结果的准确性和可靠性。结合实例验证构建模型的正确性和有效性,并与模糊数学方法及专家评价方法的结果进行对比分析,结果表明应用基于数据挖掘的集对分析方法评价地图质量是有效可行的,且操作过程简单。
The evaluation of map quality is prime importance for the improvement of map design and application. To overcome and avoid the problems of weight determination and large sample size of mining information in the traditional evaluated methods, a novel method using the are proposed to evaluate the map quality in this paper. theory of set pair analysis （SPA） based on data mining The certainty and uncertainty of objective problems are treated as uncertain systems during the process of SPA, and the uncertainties are described by the connecting degree. However, The evaluation of map quality not only need to calculate identical degree, contrary degree between the evaluated index sets of sample and grade standard sets, but also different degree and to mine useful information from the discrepancy coefficient of connecting degree. The approaching degree is introduced to describe the fuzzy connecting degree between the index values, which are located in the neighbor discussed grade, the discussed grade. Then find the maximum membership degree based on the general connecting degree of the sampie. To quantify the evaluated result, and the characteristic values are also calculated. Moreover, to confirm the reliability and the versatility of this improved SPA method, the application to a practical example, and the comparisons of evaluated results with the fuzzy mathematic method and expert evaluated method were conducted. The results show that the proposed method used to assess map quality is feasible and effective, and easy to operate. And the method has no limit on specimen modeling. This presented model also has provided the new reference for the evaluation of other uncertainty problems.
出处
《地理科学》
CSCD
北大核心
2008年第5期688-692,共5页
Scientia Geographica Sinica
基金
国家自然科学基金项目(50579009
40702049)
安徽省优秀青年科技基金项目(08040106830)资助
关键词
地图质量
集对分析
差异度系数
模糊联系度
数据挖掘
map quality
set pair analysis
discrepancy coefficient
fuzzy degree of connection
data mining
