Cross-Validation, Bootstrap, and Support Vector Machines.pdf.pdf

Hindawi Publishing CorporationAdvances in Arti?cial Neural SystemsVolume 2011, Article ID302572,6pagesdoi: ArticleCross-Validation, Bootstrap, and Support Vector MachinesMasaaki Tsujitani1, 2and Yusuke Tanaka1, 21Division of Informatics puter Sciences, Graduate School of Engineering, Osaka munication University,Osaka 572-8530, Japan2Biometrics Department, Statistics Analysis Division, EPS Co., Ltd., 3-4-30 Miyahara, Yodogawa-ku, Osaka 532-0003, JapanCorrespondence should be addressed to Masaaki Tsujitani,******@ 2 April 2011; Accepted 7 June 2011Academic Editor: Tomasz G. SmolinskiCopyright ? 2011 M. Tsujitani and Y. Tanaka. This is an open access article distributed under the mons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is paper considers the applications of resampling methods to support vector machines (SVMs). We take into account the leaving-one-out cross-validation (CV) when determining the optimum tuning parameters and bootstrapping the deviance in order tosummarize the measure of goodness-of-?t in SVMs. The leaving-one-out CV is also adapted in order to provide estimates of thebias of the excess error in a prediction rule constructed with training samples. We analyze the data froma mackerel-egg survey anda liver-disease