首页 | 本学科首页   官方微博 | 高级检索  
     检索      

地质灾害的非线性数据处理与建模技术
引用本文:许强,黄润秋.地质灾害的非线性数据处理与建模技术[J].山地学报,2000,18(Z1):123-127.
作者姓名:许强  黄润秋
作者单位:地质灾害防治与地质环境保护国家专业实验室,地质灾害防治与地质环境保护国家专业实验室 四川成都 610059,四川成都 610059
基金项目:国家自然科学基金(项目编号:49702038),国家杰出青年科学基金(项目编号:49525204)联合资助
摘    要:本文简略地介绍了几种地质灾害数据处理与建模的非线性方法,主要包括GMDH自组织建模技术、神经网络方法。GMDH是一种高阶非线性回归建模方法,它是以简单的二元二次回归方程为基础,通过"代复一代"的"生产"过程,客观、自动地求得实际资料的非线性模型。而神经网络则是用工程技术手段模拟生物神经网络的结构特征和功能特征的一类人工系统。与常规统计方法相比,神经网络最突出的优点为它是通过对网络的学习和训练,来掌握变量之间的非线性关系。因此,其处理复杂问题的能力更强大。实例检验效果表明,这些非线性数据处理与建模技术考虑了地质灾害问题的非线性特性,其比基于常规统计理论的数据处理方法的精度要高得多。

关 键 词:数据处理  数据处理的组合方法  神经网络  地质灾害  非线性科学

NONLINEAR DATA HANDLING AND MODEL ESTABLISHING OF GEOLOGICAL HAZARDS
XU Qiang,HUANG Runqiu.NONLINEAR DATA HANDLING AND MODEL ESTABLISHING OF GEOLOGICAL HAZARDS[J].Journal of Mountain Research,2000,18(Z1):123-127.
Authors:XU Qiang  HUANG Runqiu
Abstract:Two kinds of nonlinear methods of data handling and model establishing are introduced in this paper. These are group method of dada handling (abbr. GMDH ) and artificial neural network (abbr. ANN). GMDH is a higher order and nonlinear regression method. Based on the simple regression function of two variables and two orders, GMDH can get the nonlinear model by means of continual producing process automatically and objectively. Artificial neural network is one of artificial system, which simulates structural and functional characteristics of biological neural network resorted to engineering technological means. Compared with statistical and other mathematical methods, its remarkable advantage lies in that it can grasp the complicated nonlinear relationship between input and output variables by learning from given samples, and it has function of associative memory. So ANN can handle data which have very complicated relationship. The results through actual examples testing show that the two methods of GMDH and ANN are superior to routine statistical methods for complex data handling.
Keywords:data handling  group method of data handling  artificial neural network  geological hazards  nonlinear science  
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号