tdmParaBootstrap: Parametric bootstrap: add 'noisy copies' to a data frame...
In TDMR: Tuned Data Mining in R

Description Usage Arguments Value Author(s) See Also

A normal distribution is approximated from the data given in dset[,input.variables] and new data are drawn from this distribution for the columns input.variables. The column resp is filled at random with levels with the same relative frequency as in dset[,resp]. Other columns of dset are filled by copying the entries from the first row of dset.

1	tdmParaBootstrap(dset, resp, input.variables, opts)

`dset`	data frame with training set
`resp`	name of column in dset which carries the target variable
`input.variables`	vector with names of input columns
`opts`	additional parameters [defaults in brackets] `ncopies` how many noisy copies to add `ncsigma` [1.0] multiplicative factor for each std.dev. `ncmethod` [3] which method to use for parametric bootstrap =1: each 'old' record from X in turn is the centroid for a new pattern; =2: the centroid is the average of all records from the same class, the std.dev. is the same for all classes; =3: centroid as in '2', the std.dev. is the std.dev. of all records from the same class (recommended) `TST.COL` (optional) name of column in dset where each PB record is marked with a 0