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

ANFIS与SVM的拟合能力和推广预测能力的比较研究
引用本文:张志军,丁德馨,贺桂成.ANFIS与SVM的拟合能力和推广预测能力的比较研究[J].南华大学学报(自然科学版),2009,23(4):14-19.
作者姓名:张志军  丁德馨  贺桂成
作者单位:南华大学,核资源与核燃料工程学院,湖南,衡阳,421001
基金项目:2006年教育部高校博士点专项基金资助项目,湖南省自然科学基金联合基金资助项目,湖南省教育厅基金资助项目,国家安全监管总局2007年度安全生产科技发展计划基金资助项目,湖南省安全生产科技发展指导性计划基金资助项目 
摘    要:岩土体位移与岩土体力学参数间的映射关系,具有高度非线性的特点.寻求一种适合于描述这种非线性映射关系的方法,是当前的热点研究领域.为此,本文采用一个多峰函数进行离散,构建了训练数据对和预测数据对,分别对ANFIS和SVM的拟合能力和推广预测能力进行了比较研究,结果表明,ANFIS的拟合能力和推广预测能力均优于SVM,更适合于建立岩土体位移与岩土体力学参数间这一高度非线性映射关系.

关 键 词:支持向量机  自适应神经模糊推理方法  推广预测能力
收稿时间:2009/10/23 0:00:00

A Comparison Study of the Fitting and Generalization Prediction Capabilities of SVM and ANFIS
ZHANG Zhi-jun,DING De-xin,HE Gui-cheng.A Comparison Study of the Fitting and Generalization Prediction Capabilities of SVM and ANFIS[J].Journal of Nanhua University:Science and Technology,2009,23(4):14-19.
Authors:ZHANG Zhi-jun  DING De-xin  HE Gui-cheng
Institution:School of Nuclear Resources and Nuclear Fuel Engineering,University of South China,Hengyang,Hunan 421001,China
Abstract:The mapping between displacements and mechanical properties of rock and soil mass is characterized by its non-linearity.Many researchers are trying to establish an approach for describing this non-linear mapping.As a result,data pairs for training and data pairs for prediction were built by using a multi-peaks function,and a comparison study was conducted for the fitting and generalization prediction capabilities of ANFIS and SVM.The results show that ANFIS has better fitting and generalization prediction capabilities than SVM and it adapts to deal with the nonliner and complex mapping problem in geotechnique engineering.
Keywords:support vector machine  ANFIS  generalization prediction capability
本文献已被 维普 万方数据 等数据库收录!
点击此处可从《南华大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《南华大学学报(自然科学版)》下载免费的PDF全文
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