中文摘要：

云南省是我国各种地质灾害的多发区。该文以2000—2014年云南省发生的500个典型地质灾害事件为基础，选取与地质灾害相关的诱发因素因子和潜在形成条件因子，构建了相应的评价指标体系。然后采用地理加权回归模型对其进行易发性评价，并对结果进行了等级划分。最后利用引入空间准确测量思想的成功率曲线和ROC曲线，对评价结果进行了精度验证和分析比较。结果表明：（1）云南省地质灾害较高和极高易发区主要位于西北部地区，且以泸水县、中甸县和维西傈僳族自治县为最高；（2）对评价结果进行了分级，从低到高不同易发性等级的面积百分比分别为6．14％，24．80％，18．41％，24．68％，20．78％和5．18％；（3）在未引入空间准确测量思想情况前，成功率曲线和ROC曲线的AUC值分别为0．5038和0．9651；引入空间准确测量思想后，AUC值分别为0．6813和0．9927，证明所提基于地理加权回归模型的云南省地质灾害易发性评价具有较高的可靠性。

英文摘要：

Yunnan Province is a geological disaster-prone region in China. In this paper, we collected 500 typical geological disasters events occurred in Yunnan Province in 2000 -2014 and data set of their induced factors, as well as potential formation conditions and constructed corresponding evaluation index system. Then we used geo- graphically weighted regression（GWR） model to evaluate susceptibility of the geological disasters, and divided the results into some levels. Finally, we utilized the success rate curve introducing space accurate measurement ideas and ROC curve to carry out accuracy verification and analytical comparison. The results show that： （ 1 ） the most and more susceptible area of geological disasters in Yunnan Province is mainly located in its northwest region, such as Lushui County, Zhongdian County and Weixi Lisu Autonomous County. （2） The area percentages of different susceptibility levels from low to high are 6.14%, 24.80% , 18.41% , 24.68%, 20.78% and 5.18%. （3）Be- fore introducing space accurate measurement thought, the AUC indexes of success rate curves and ROC curve are 0. 503 8 and 0. 965 1 ; and after that, they are 0. 681 3 and 0.992 7 respectively, which proves that GWR model isreliable in evaluating the susceptibility of geological disasters for Yunnan Province.