pcaKern: Kernel Principal Components Analysis
In abnormally-distributed/cvreg: Cross Validation and Robust Estimation Utilities

Kernel Principal Components Analysis

1	pcaKern(x = NULL, ncomp = NULL, scale = T, kern = RbfKernel, ...)

`x`	a data frame or matrix of numeric variables. can be NULL if a kernel matrix is supplied directly to argument 'kern'.
`ncomp`	specify the maximum number of components to be returned. if there are fewer components with non-zero eigenvalues than the number specified, then all components with non-zero eigenvalues will be returned. at least two components will be returned regardless of what is specified here.
`scale`	should the variables be scaled prior to analysis? Defaults to TRUE.
`kern`	either a kernlab compatible kernel function, i.e., 'laplacedot(1)', or a function that generates square, symmetric, kernel matrix. a kernel matrix can also be directly supplied if one has been created. centering is done internally so it is not necessary to pre-center the kernel matrix. The default is the radial basis function kernel with the scale parameter automatically determined. Unless the default gives poor results, the rbf kernel with the automatic scale parameter is safe to leave as-is.
`...`	additional arguments to supply if the argument 'kern' is given a function that generates a matrix (ie, not a kernlab function). anything supplied here will be passed to the function, so the argument must match one of the parameters in the kernel.