The number of *best* variables is estimated by finding an
apparent inflection point in the relationship between the
generalized root mean square distance (see `grmsd`

and the number
of *X*-variables.

1 |

`obj` |
an object create by |

`nbest` |
number of variables designated as the best; if null the number is estimated |

An character vector of variable names in decreasing order of importance.

Nicholas L. Crookston ncrookston.fs@gmail.com

`varSelection`

1 2 3 4 5 6 7 8 9 10 11 |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

All documentation is copyright its authors; we didn't write any of that.