CVST-package: Fast Cross-Validation via Sequential Testing

In CVST: Fast Cross-Validation via Sequential Testing

Fast Cross-Validation via Sequential Testing

Description

The fast cross-validation via sequential testing (CVST) procedure is an improved cross-validation procedure which uses non-parametric testing coupled with sequential analysis to determine the best parameter set on linearly increasing subsets of the data. By eliminating under-performing candidates quickly and keeping promising candidates as long as possible, the method speeds up the computation while preserving the capability of a full cross-validation. Additionally to the CVST the package contains an implementation of the ordinary k-fold cross-validation with a flexible and powerful set of helper objects and methods to handle the overall model selection process. The implementations of the Cochran's Q test with permutations and the sequential testing framework of Wald are generic and can therefore also be used in other contexts.

Details

This package was not yet installed at build time.
Index: This package was not yet installed at build time.

Author(s)

Tammo Krueger, Mikio Braun

Maintainer: Tammo Krueger <[e-mail]>

References

Tammo Krueger, Danny Panknin, and Mikio Braun. Fast cross-validation via sequential testing. Journal of Machine Learning Research 16 (2015) 1103-1155. URL https://jmlr.org/papers/volume16/krueger15a/krueger15a.pdf.

Abraham Wald. Sequential Analysis. Wiley, 1947.

W. G. Cochran. The comparison of percentages in matched samples. Biometrika, 37 (3-4):256–266, 1950.

M. Friedman. The use of ranks to avoid the assumption of normality implicit in the analysis of variance. Journal of the American Statistical Association, 32 (200):675–701, 1937.

Examples

ns = noisySine(100)
svm = constructSVMLearner()
params = constructParams(kernel="rbfdot", sigma=10^(-3:3), nu=c(0.05, 0.1, 0.2, 0.3))
opt = fastCV(ns, svm, params, constructCVSTModel())

CVST documentation built on March 18, 2022, 5:59 p.m.