文献类型：期刊文献

中文题名：滑坡成因决策树挖掘

英文题名：Decision-making tree for analysis of landslide causes

作者：亓呈明[1];崔守梅[2];陈辉[3];曾青石[3]

第一作者：亓呈明

机构：[1]北京联合大学自动化学院;[2]山东省淄博师范高等专科学校;[3]中国地质环境监测院

第一机构：北京联合大学城市轨道交通与物流学院

年份：2006

卷号：17

期号：1

起止页码：73-76

中文期刊名：中国地质灾害与防治学报

外文期刊名：The Chinese Journal of Geological Hazard and Control

收录：CSTPCD;;CSCD:【CSCD_E2011_2012】；

语种：中文

中文关键词：数据挖掘;决策树;滑坡

外文关键词：data mining; decision tree ; landslide

摘要：文章介绍了构造决策树的C4.5算法。在预处理过的滑坡数据上利用决策树算法生成一组坡体稳定性规则。这些规则可作为分析及预测滑坡的重要参考。也可用来创建专家系统的知识库。其次,作者对决策树算法提出了改进,通过调整测试集的某些参数,构造了一棵局部最优的决策树。使得对于滑坡成因的分类具有更高的正确率。
This paper introduces C4.5 algorithm, which is widely used to construct decision-making tree. By implementation of this algorithm, a set of rules of slope stability can be generated based on a series of preprocessed landslide data. The generated rules could be subsequatly used for landslide analysis and predction, also could be used in building of repository of landslide expert system. To construct an optimized decision-making tree. this algorithm was further modified by adjusting same parameters in test data set. The modified algorithm can increase correction ratio when it is used for classification of landslide causes.