基于粗集决策树的混凝土坝裂缝成因挖掘  

Cause Mining of Crack in Concrete Dam Based on Rough Sets and Decision Tree

作　　者：薛凌峰[1,2] 李靖[1,2] 

机构地区：[1]河海大学水利水电工程学院,江苏南京210098 [2]水资源高效利用与工程安全国家工程研究中心,江苏南京210098

出　　处：《水力发电》2008年第11期45-47,60,共4页Water Power

基　　金：国家自然科学基金项目(50579010);国家科技支撑计划项目(20006BAC14BO3);国家自然科学基金项目(50539110);国家自然科学基金重点项目(50539010);中国水电工程顾问集团公司科技项目(CHC-KJ-2007-02);国家自然科学基金(50539030-1-3)

摘　　要：在收集、提取与混凝土坝裂缝成因有关数据和信息的基础上,采用粗集理论和决策树相结合的数据挖掘方法,进一步挖掘混凝土坝裂缝产生和发展的主要原因。首先运用粗集理论对条件属性约简,去除冗余属性,然后采用C4.5算法构建决策树,并对决策树剪枝提取规则知识。实例表明,采用粗集和决策树相结合的方法挖掘混凝土坝裂缝成因是可行的。On the basic of collecting and extracting the information related to the cracks in concrete dam, the data mining method which combines rough sets with decision tree is further used to mine the main causes of production and development of cracks in concrete dam. Firstly, rough sets is used to reduce condition attribute and remove redundancy attribute, and then, C4.5 is used to build decision tree, rule knowledge is extracted by decision tree pruning. The example of engineering shows that it is feasible to use the data mining method which combines rough sets with decision tree to mine the main causes of crack in concrete dam.

关 键 词：数据挖掘 粗集理论 属性约简 决策树 C4.5算法 PEP 裂缝成因 混凝土坝 

分 类 号：TV642[水利工程—水利水电工程]