Rinverse: Difference between revisions
Jump to navigation
Jump to search
imported>Jeremy (Importing text file) |
imported>Chuck No edit summary |
||
| (3 intermediate revisions by 2 users not shown) | |||
| Line 1: | Line 1: | ||
===Purpose=== | ===Purpose=== | ||
Calculates pseudo inverse for PLS, PCR and RR models. | Calculates pseudo inverse for PLS, PCR and RR models. | ||
===Synopsis=== | ===Synopsis=== | ||
:rinv = rinverse(mod,ncomp) | |||
:rinv = rinverse(p,t,w,ncomp) | :rinv = rinverse(mod,ncomp) % for PLS or PCR model structures | ||
:rinv = rinverse(p,t,ncomp) | :rinv = rinverse(p,t,w,ncomp) % for PLS | ||
:rinv = rinverse(sx,theta) | :rinv = rinverse(p,t,ncomp) % for PCR | ||
:rinv = rinverse(sx,theta) % for RR | |||
===Description=== | ===Description=== | ||
This function calculates the pseudo-inverse of a PLS, PCR or Ridge Regression (RR) model, using either a standard model structure, or a series of model parameter arrays. There are four different input cases, depending on the type of model, and whether a standard model structure exists: | |||
For PLS models, the inputs are the loadings p, scores t, weights w and number of | ====Input Cases==== | ||
rinv = rinverse(p,t,w,ncomp) | |||
For PCR models, the inputs are the loadings p, scores t, and number of | * For PLS and PCR model structures, the inputs are the model <tt>mod</tt> and the optional number of factors <tt>ncomp</tt>. | ||
rinv = rinverse(p,t,ncomp) | |||
For ridge regression (RR) models, the inputs are the scaled | :rinv = rinverse(mod,ncomp); | ||
rinv = rinverse(sx,theta) | |||
* For PLS models, the inputs are the loadings <tt>p</tt>, scores <tt>t</tt>, weights <tt>w</tt> and number of LVs <tt>ncomp</tt>. | |||
:rinv = rinverse(p,t,w,ncomp); | |||
* For PCR models, the inputs are the loadings <tt>p</tt>, scores <tt>t</tt>, and number of PCs <tt>ncomp</tt>. | |||
:rinv = rinverse(p,t,ncomp); | |||
* For ridge regression (RR) models, the inputs are the scaled x matrix <tt>sx</tt> and ridge parameter <tt>theta</tt>. | |||
:rinv = rinverse(sx,theta). | |||
====Outputs==== | |||
* '''rinv''' = pseudo-inverse of model | |||
===See Also=== | ===See Also=== | ||
[[pcr]], [[pls]], [[ridge]], [[stdsslct]] | [[pcr]], [[pls]], [[ridge]], [[stdsslct]] | ||
Latest revision as of 11:24, 9 October 2008
Purpose
Calculates pseudo inverse for PLS, PCR and RR models.
Synopsis
- rinv = rinverse(mod,ncomp) % for PLS or PCR model structures
- rinv = rinverse(p,t,w,ncomp) % for PLS
- rinv = rinverse(p,t,ncomp) % for PCR
- rinv = rinverse(sx,theta) % for RR
Description
This function calculates the pseudo-inverse of a PLS, PCR or Ridge Regression (RR) model, using either a standard model structure, or a series of model parameter arrays. There are four different input cases, depending on the type of model, and whether a standard model structure exists:
Input Cases
- For PLS and PCR model structures, the inputs are the model mod and the optional number of factors ncomp.
- rinv = rinverse(mod,ncomp);
- For PLS models, the inputs are the loadings p, scores t, weights w and number of LVs ncomp.
- rinv = rinverse(p,t,w,ncomp);
- For PCR models, the inputs are the loadings p, scores t, and number of PCs ncomp.
- rinv = rinverse(p,t,ncomp);
- For ridge regression (RR) models, the inputs are the scaled x matrix sx and ridge parameter theta.
- rinv = rinverse(sx,theta).
Outputs
- rinv = pseudo-inverse of model