fwdmv.object {Rfwdmv}R Documentation

fwdmv.object

Description

An object containing a fitted forward search on multivariate data. The class attribute is set to fwdmv.

Format

An fwdmv.object is a list with the following elements:

call
the matched call.
Distances
a numeric matrix containing the Mahalanobis distances computed during the forward search.
Center
a list of numeric matrices containing the location estimates for each group computed during the forward search.
Cov
a list of numeric matrices containing the covariance matrix estimates (in packed storage) for each group computed during the forward seach).
Determinant
a list of numeric vectors containing the determinants of the covariance matrix estimates for each group computed during the forward search.
Unit
a matrix with one row for each unit in the data set and one column for each step in the forward search. The (i,j) element is 1 if unit i is in the subset during step j of the forward search and is 0 otherwise.
groups
a list of integer vectors containing the user specified tentative groups.
n
an integer, the number of units in the data.
p
an integer, the number of variables in the data.
m
an integer, the number of units in the subset during the first step of the forward search.
data
a numeric matrix containing the data, the dimnames attribute is set to NULL.
data.name
the name of the data frame or matrix containing the multivariate data set.
data.names
a list of character vectors containing the row and column names of the data.
group.names
a character vector containing the names of the tentative groups.
unassigned
an integer vector containing the indices of the units that do not belong to one of the tentative groups.
constrained
a loigcal value, TRUE if the forward search was constrained.
scaled
a logical value, TRUE is scaled Mahalanobis distances were used during the forward search.
Max
a numeric vector contining the maximum Mahalanobis distance in the subset.
Mth
a numeric vector containing the mth overall Mahalanobis distance.
Min
a numeric vector containing the minimum Mahalanobis distance not in the subset.
Mpo
a numeric vector containing the (m+1)th overall Mahalanobis distance.
initial
a logical value, TRUE if this object was generated by the function fwdmv.init.

Details

Class fwdmv objects are created by the functions fwdmv, fwdmv.init, and partition. The Rfwdmv package contains a variety of plot methods for assesing fwdmv objects. These methods are listed in the see also section. Note that there are accessor methods for several elements contained in the fwdmv object. When an accessor method exists for a certain element it should be used to retrieve that element in preference to direct reference as the structure of the fwdmv object is likely to change.

See Also

fwdmv.init
fit an ititial multivariate forward search.
fwdmv
fit a multivariate forward search with user specified tentative groups.

Plot methods for assessing a fitted forward search stored in an fwdmv object:

fwdmvPairsPlot
a pairs-like plot.
fwdmvQuantilePlot
plot trajectories of entering units over quantiles of the distances in the subset.
fwdmvEllipsePlot
a pairs-like plots with the subsets represented by ellipses.
fwdmvConfirmPlot
plots the nearest center for unassigned and misclassified units.
fwdmvCovariancePlot
a forward plot of the elements of the covariance matrices.
fwdmvDeterminantPlot
a forward plot of the determinants of the covariance matrices.
fwdmvDistancePlot
a forward plot of the Mahalanobis distances.
fwdmvEccentricityPlot
a forward plot of the eccentricity for one biariate ellipse.
fwdmvPrincompPlot
a forward plot of the principal components.
fwdmvEigenvectorPlot
a forward plot of a user specified eigenvector of the covariance matrice.
fwdmvEntryPlot
a forward entry plot.
fwdmvGapPlot
a gap plot.
fwdmvMinmaxPlot
minimum and maximum distances plot.
fwdmvChangePlot
aa forward plot of change in Mahalanobis distance.
partition
graphically assign units to groups.
fwdmvPartitionPlot
view tentative group.

Accessor methods:

eigenvalues.fwdmv
the eigenvalues of the covariance matrix estimates computed during the forward search.
eigenvectors.fwdmv
the eigenvectors of the covariance matrix estimates computed during the forward search.

Examples

data(fondi.dat)
fondi.1 <- fwdmv(fondi.dat)

#fondi.1 is an fwdmv object.

[Package Rfwdmv version 0.72-2 Index]