Variable selection methods are provided for several classification methods: the lasso/elastic net, PCLDA, PLSDA, and several t-tests. Two approaches for selecting cutoffs can be used, one based on the stability of model coefficients under perturbation, and the other on higher criticism.
|Author||Ron Wehrens, Pietro Franceschi|
|Date of publication||2015-09-07 18:17:37|
|Maintainer||Ron Wehrens <email@example.com>|
|License||GPL (>= 2)|
auxBiom: Auxiliary functions in the biomarker package
biom.options: Set or return options for stability-based biomarker selection
gen.data: Simulate data sets
get.biom: Get biomarkers discriminating between two classes
get.segments: Subsampling segments
HCthresh: Biomarker thresholding by Higher Criticism
ROC: ROC curves
scalefun: Different forms of scaling
selection: Accessor function to the selected variables of a BioMark...
SpikedApple: Spike-in metabolomics data for apple extracts
spikedApples: Metabolomics data on spiked apples
traceplot: Plot the coefficient or stability trace for the lasso/elastic...