系统加权聚类法及其在滑坡稳定性预测中的应用  

A HIERARCHICAL WEIGHTED CLUSTERING METHOD AND ITS APPLICATION IN PREDICTING LANDSLIDE STABILITY

作　　者：夏元友[1] 

机构地区：[1]武汉工业大学岩土与环境工程研究所

出　　处：《自然灾害学报》1997年第3期85-91,共7页Journal of Natural Disasters

摘　　要：本文以工程地质类比分析法为基础，针对一般系统聚类法在滑坡稳定性预测中应用的不足，提出并建立了一种系统加权聚类法。最后以三峡库岸研究程度较高的３６个边坡为例，进行了一般系统聚类法与系统加权聚类法的对比研究。对比研究结果表明：系统加权聚类法应用于滑坡稳定性预测，其预测精度有较大提高。This paper presents a hierarchical weighted clustering method (HWCM) for predicting landslide stability based on engineering geology analogy analysis method,and this method eliminates the defects of traditional hierarchical clustering method (HCM) which think that all evaluation indexes (effectived indexes of landslide stability) are of the same importance.The paper also uses HWCM and HCM separately to classify the stabilities of 36 slopes on the banks of the Three Gorges Reservoir.The comparison analysis shows that HWCM has higher precision than HCM in predicting slope stability.

关 键 词：灾害预测 滑坡 系统加权 聚类分析法 稳定性 

分 类 号：P642.22[天文地球—工程地质学]