<< pls (X->y) Linear calibrations ridge >>

fact >> (X->y) Linear calibrations > regapply

regapply

applies a calibration model calculated with pls, vodka, pcr, mlr or ridge to a new dataset

Calling sequence

res=regapply(model,x,(y),(plot))

Arguments

model:

model obtained with one of the regression methods (pls, vodka, pcr, mlr or ridge)

x , (y) :

test dataset; a matrix (n x q) and an optional vector of reference values y of dimensions (n x 1), or Div structures

(plot):

figure representing y and ypredtest as abscissa and ordinate respectively; R2, biais and RMSEP are also calculated

the figure is plotted only if the field 'plot' is not empty

res.ypredtest:

predicted values

ypred.d is a matrix (n x lv) with lv = number of calibration models in the 'model' structure

(res.rmsep):

root mean square error of prediction (if y is given)

rmsep.d is a vector (lv x 1)

(res.r2):

squared correlation coefficient (if y is given)

r2.d is a vector (lv x 1)

Examples

[res1]=regapply(plsmodel,xtest)
[res2]=regapply(plsmodel,xtest,ytest,1)

Authors


Report an issue
<< pls (X->y) Linear calibrations ridge >>