ClassExReg: Classification Extrapolation with Regression (and bootstrap...
In snarles/ClassEx: Extrapolation Methods for Multiclass Classification Accuracy

Classification Extrapolation with Regression (and bootstrap model selection)

1 2	ClassExReg(pmat, i_chosen, basis_sets, Ktrain = 2:ncol(pmat), nboot = 25, nonnegative = TRUE)

`pmat`	Matrix of margins m_y(x) for test instances, rows are observations and columns are labels.
`i_chosen`	True labels.
`basis_sets`	a list of bases obtained e.g. using lapply and get_gaussian_basis_mat. If there is only one element in the list, then model selection is not applied.
`Ktrain`	the number of classes k for which we have computed average test accuracies ATA_k.
`nboot`	Number of bootstrap samples to use in model selection.
`nonnegative`	constrain coefficients of basis to be nonnegative? default TRUE.