Research on Response Surface Method of Support Vector Machine based on Markov Chain

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The common form of response surface methods affect the accuracy of reliability calculation in the experimental design because of the over dependence on the selection of experimental points. Based on the characteristics of the Least Square Support Vector Machine (LSSVM), a method to obtain the approximate designed points was designed by using an adaptive Markov Chain to simulate the samples in the failure domain and the safety domain. The improved selection method of design point could be used to select the samples adaptively at the real limit state boundary, to improve the fitting precision of the real boundary and increase the calculation precision of reliability. In this article, two cases of the multiple failure mode and the single failure mode with complex boundary are studied and compared to other methods to illustrate the advantages of the proposed method.

Original languageEnglish
Title of host publicationProceedings of 2018 IEEE 3rd Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2018
EditorsBing Xu
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages2359-2364
Number of pages6
ISBN (Electronic)9781538645086
DOIs
Publication statusPublished - 14 Dec 2018
Event3rd IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2018 - Chongqing, China
Duration: 12 Oct 201814 Oct 2018

Publication series

NameProceedings of 2018 IEEE 3rd Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2018

Conference

Conference3rd IEEE Advanced Information Technology, Electronic and Automation Control Conference, IAEAC 2018
Country/TerritoryChina
CityChongqing
Period12/10/1814/10/18

Keywords

  • markov chain
  • reliability
  • response surface method
  • support vector machine

Fingerprint

Dive into the research topics of 'Research on Response Surface Method of Support Vector Machine based on Markov Chain'. Together they form a unique fingerprint.

Cite this