Abstract
TIMP is an R package for modeling multiway spectroscopic measurements. The package allows for the simultaneous analysis of datasets collected under different experimental conditions in terms of a wide variety of parametric models. Models arising in spectroscopy data analysis often have some parameters that are intrinstically nonlinear, and some parameters that are conditionally linear on estimates of the nonlinear parameters. TIMP fits such separable nonlinear models using partitioned variable projection, a variant of the variable projection algorithm that is described here for the first time. The of the partitioned variable projection algorithm allows fitting many models for spectroscopy datasets using much less memory as compared to under the standard variable projection algorithm that is implemented in nonlinear optimization routines (e.g., the plinear option of the R function nls), as is shown here. An overview of modeling with TIMP is also given that includes several case studies in the application of the package.
Highlights
TIMP1 is an R package for modeling multiway spectroscopic measurements
The of the partitioned variable projection algorithm allows fitting many models for spectroscopy datasets using much less memory as compared to under the standard variable projection algorithm that is implemented in nonlinear optimization routines, as is shown here
S4 classes defined by TIMP represents single dataset Ψq and an associated model kinetic model specification, inherits from dat spectral model specification, inherits from dat parameter estimates associated with model for single dataset Ψq parameter estimates associated with multiple dataset model all information associated with multiple dataset model and fit results of fitting many datasets fixed parameters, parameter relations, and parameter constraints is performed in getThetaCl, not the residual function, further facilitating its implementation and readability
Summary
TIMP1 is an R package for modeling multiway spectroscopic measurements It is a fully crossplatform problem-solving environment for fitting a wide range of models to spectroscopy data, and is freely available under the GNU General Public License (GPL). TIMP: Modeling Multi-way Spectroscopic Measurements in R cludes a description of the partitioned variable projection algorithm that allows application of the variable projection functional to parameter estimation problems in the absence of large memory resources.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have