基于粗糙集的范例推理在泥石流危险性评价中的应用  

Application of Case-based Reasoning Method to Assess Debris Flow Hazard Based on Rough Set

作　　者：刘勇健[1] 刘义建 

机构地区：[1]广东工业大学岩土工程研究所,广东广州510006 [2]湖南省地质矿产勘查开发局407队,湖南怀化418000

出　　处：《山地学报》2008年第2期223-229,共7页Mountain Research

基　　金：广东省自然科学基金项目(6021462);广东省岩土工程重点学科基金资助~~

摘　　要：将粗糙集理论和范例推理相结合,建立了基于粗糙集-范例推理的泥石流危险性评价模型。运用粗糙集理论对范例库中的范例进行属性约简和特征向量权值计算,用相似度量理论来实现最相似范例检索,得到泥石流的危险性评价结果。选取了一次泥石流(可能)最大冲出量,泥石流发生频率、流域面积、主沟长度、流域相对高差、流域切割密度、24h最大降雨量、松散固体物质储量共8个因素作为泥石流危险性的影响因子;选取云南东川和西藏八宿等共14条泥石流沟资料进行建模分析。实例研究表明,基于粗糙集-范例推理的泥石流危险性评价结果与实际状态相吻合,该方法具有简便、高效、直观、实用的特点。To assess debris flow hazard, A case-based reasoning model （CBR）, combining with rough set, is established in the paper. The rough set is used to reduce the features of the case and calculate features weights, while the similarity degree measurement, which improves retrieval speed and accuracy, is used to retrieve the similar cases. Thus the above methods can obtain the criticality of debris flow. In the paper, eight factors including the maximum expulsion amount of debris flow once, occurrence frequency, drainage area, length of the main gully, relative altitude, cutting density, maximum quantity of rainfall in 24 hours and amount of loss solid matter, are chosen as the influence factors of debris flow hazard. 14 debris flows of Dongchuan in Yunnan Province and Basu in Tibet are ap- plied to set up model of debris flow hazard assessment. The research shows that the results of debris flow hazard assessment are the same as the practical states, and this method is simple, effective, visual and practical.

关 键 词：泥石流 危险性评价 粗糙集 范例推理 相似度量 

分 类 号：P642.23[天文地球—工程地质学]