TIMP-package {TIMP} | R Documentation |
Measurements often represent a superposition of the contributions of distinct sub-systems resolved with respect to many experimental variables (time, temperature, wavelength, pH, polarization, etc). A parametric model for each component may be desirable to apply to the data, but only to the evolution of components with respect to a subset of the independent variables. For instance, given time-resolved spectroscopy data, a parametric model for the time-evolution of components may be available, while a physically-inspired parametric model for the spectra of components may be difficult to formulate and interpret. Such situations give rise to a separable nonlinear parameter estimation problem, namely that of estimating the (nonlinear) parameters associated with the parametric model, while estimating parameters representing the evolution of components with respect to the independent variables to which a parametric model does not apply as conditionally linear. The partitioned variable projection algorithm is well-suited to solving such problems under the criteria of efficiency, quality of standard error estimates, and precision of parameter estimates. TIMP implements the partitioned variable projection algorithm and allows its application to fitting a wide range of models, including those for the simultaneous analysis of multiple datasets collected under different experimental conditions. The package has been extensively applied to modeling data arising in spectroscopy experiments.
Package: TIMP Type: Package Title: a problem solving environment for fitting superposition models Version: 1.2 Author: Katharine M. Mullen, Ivo H. M. van Stokkum Maintainer: Katharine M. Mullen <kate@nat.vu.nl> Depends: R (>= 2.5.0), methods, tcltk, vcd, fields, gplots, splines Suggests: gclus License: GPL version 2 or newer
Katharine M. Mullen, Ivo H. M. van Stokkum Maintainer: Katharine M. Mullen kate@nat.vu.nl
See http://www.nat.vu.nl/~kate/TIMP/ for further documentation.