摘要
泥石流是自然及人类活动共同作用影响下的一种山区地质灾害,具有突发性、流速快、流量大及破坏力强等特点,泥石流易发性评价是灾害预测预报和减灾防灾工作中的重要内容。基于因子分析法将影响泥石流易发性等级的指标降维提取几个具有代表性的公因子,将因子得分进行欧氏聚类分析,剔除相似度较低的错误样本,将剩余具有相似的基本特征的几种同类型泥石流,根据Bayes最大后验概率规则,创建基于多元统计Bayes判别原理的泥石流易发性评价模型,并对其易发性等级判别结果的准确性进行了验证。结果表明:将唐山北部地区12条泥石流指标的公因子得分代入判别式中,得到泥石流易发性等级与实际等级对比结果一致,表明基于因子分析和聚类分析两种多元统计分析方法建立的多元统计Bayes判别模型是合理可行的,评价结果与实际情况具有较好的一致性,为泥石流易发性研究提供了一种新的研究手段。
Debris flow is a kind of mountain geological disaster under the influence of natural and human activities.It has the characteristics of sudden,fast flow,large flow and strong destructive power.The evaluation of debris flow vulnerability is an important part of disaster prediction and mitigation.This article is based on the factor analysis method,several representative common factors are extracted from the dimension reduction of the indicators that affect the vulnerability level of debris flow.The factor scores are analyzed by Euclidean clustering,the wrong samples with low similarity are removed,and the remaining debris flows with similar basic characteristics are classified into the same type.According to Bayes maximum posterior probability rule,a debris flow vulnerability evaluation model based on multivariate statistical Bayes discrimination principle is created.The accuracy of the results of the susceptibility level discrimination was verified.The results show that the common factor scores of the 12 debris flow indicators in the northern Tangshan area are substituted into the discriminant,and the comparison results of the debris flow vulnerability grade and the actual grade are consistent.It shows that the multivariate statistical Bayes discriminant model based on the two multivariate statistical analysis methods of factor analysis and cluster analysis is reasonable and feasible,and the evaluation results are consistent with the actual situation.It could provides a new research method for the study of debris flow vulnerability.
作者
单强
汪洋
王颖
马丙太
王蒙
SHAN Qiang;WANG Yang;WANG Ying;MA Bingtai;WANG Meng(Hebei Geo-environment Monitoring,Shijiazhuang 050021,China;Hebei Key Laboratory of Geological Resources and Environment Monitoring and Protection,Shijiazhuang 050021,China;Hebei Institute of Geological Survey,Shijiazhuang 050081,China)
出处
《甘肃科学学报》
2024年第3期122-129,共8页
Journal of Gansu Sciences
关键词
泥石流
易发性
多元统计
Bayes判别
Debris flow
Susceptibility
Multivariate statistics
Bayes discrimination
