xxx

1 2 | ```
relCurvature(s.sqr, R11, R2, show.extra.details=FALSE,
eps=sqrt(.Machine$double.eps))
``` |

`s.sqr` |
sample estimate of the residual variance. |

`R11` |
submatrix of R. See details. |

`R2` |
submatrix of R. See details. |

`show.extra.details` |
logical indicating if extra intermediate calculations should be printed. |

`eps` |
singular values smaller than eps times the largest singular value are considered to be zero. |

The result is the relative curvature array Bates & Watts p242-244 eqn. (7.16) and examples p244-245.

R11 is a p by p sub matrix of R. See Bates & Watts p236 and R2 is the m by pp sub matrix R12 R22 ... If I-B is not positive definite, where B is the effective residual curvature matrix, and warn is TRUE, then a warning will indicate that the calculation does not seem to correspond to a local minimum.

relative curvature array

`curvature`

`project`

`effectiveCurvature`

`curvatureStats`

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.