An S4 class container for the result of the function
calcPLSPCE
.
indexes
:matrix with as many rows as inputs and three columns. Values of the PLS-PCE sensitivity indexes. The column labels are LE, PE, TPE.
indexes[i, "LE"]
is the Linear Effect of the input i
.
indexes[i, "PE"]
is the Polynomial Effect. It is the effect of the
monomials in which only the input i
appears.
indexes[i, "TPE"]
is the Total Polynomial Effect. It is the effect of all the
monomials in which the input i
appears.
indexes.percent
:matrix.
Percentages of the PLS-PCE sensitivity indexes, i.e
values of indexes
expressed as percentages
of the sums of their columns.
ncopt
:number of the optimal component.
R2
:matrix with as many rows as components and three columns. The columns are labeled R2, %R2 and %R2cum. They store respectively the R2 values, percentages of their column sums, and cumulated values of their percentages.
Q2
:matrix with as many rows as components and two columns. The columns are labeled Q2 and Q2cum. They store respectively the Q2 and the Q2cum values.
rmsep
:matrix with as many rows as components and one column. Root mean square error predictions.
COEF
:matrix with as many rows as monomials plus one and as many columns as components. PLS-regression coefficients. The first row is the constant term.
betaCR
:matrix with as many rows as monomials and as many columns as components. The centered-reducted PLS-regression coefficients, for all the components.
y.hat
:vector of length equal to the number of rows of the dataset. Metamodel outputs.
STRUC
:object of class PCEdesign
coding the polynomial structure.
signature(x = "PLSPCE", all=FALSE, ...)
:
method of function print
.
When option all
is FALSE, some
slots are hidden :
RMSEP, coef, betaCR, y.hat, STRUC
.
When option all
is set to TRUE,
the names of the hidden slots are printed. The additional arguments are passed to the
print.default
function.
signature(object = "PLSPCE")
: same as
function print
, without any arguments.
signature(x = "PLSPCE", pce, options =c("fit", "bar", "compo"))
:
method of function plot
.
Different plots are drawn, according to options
.
When it includes “fit”,
computer model outputs against metamodel outputs.
When it includes “bar”,
barplots of the polynomial
and total polynomial effects (PE and TPE).
When it includes “compo”,
TPE against components. This last plot requires the
calculation of the TPE for each component, which may take some time.
The COEF
and STRUC
slots can be used for making
predictions
for Legendre-coded new input values.
Function calcPLSPCE
, creator of objects from this
class.
Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.
Please suggest features or report bugs with the GitHub issue tracker.
All documentation is copyright its authors; we didn't write any of that.