attvalues: Plot distributions of selected attributes within a specified...

In mkampert/rCOSA: Clustering Objects on Subsets of Attributes (COSA)

Description

This procedure plots a matrix of histograms. Each column pertains to one of the selected attributes. The first row shows the distribution for those objects within the specified group. The second row plots the corresponding distribution for those objects not in the specified group. Optionally (all= TRUE), a third row of histograms is produced showing the distribution of the corresponding attribute over all of the data. In all histograms, only nonmissing values of the respective attributes are plotted.

Usage

attvalues(
  X,
  group,
  atts,
  lX = NULL,
  att.names = 1:ncol(X),
  group.name = "Group",
  xmiss = NULL,
  nbins = 25,
  targ = NULL,
  all = F
)

Arguments

X	input data. Either a data.frame or a matrix in mode numeric. Rows are objects and columns are attributes (variables).
group	vector of identities of objects in the group.
atts	vector of identities of selected attributes.
lX	vector of attribute flags (see cosa2).
att.names	vector of names for each attribute.
group.name	name for selected group (cluster).
xmiss	missing value flag (see cosa2).
nbins	(approXimate) number of histogram bins.
targ	target values for computing targeted dissimilarities. The targ[k] is the target value for X[,k]. The value is ignored if lX[k] = 0, 1, or 4. If lX[k]=2 or 5, then targ[k] contains the single target value. If lX[k]=3 or 6, then targ[k] contains one of the two target values. Special values: targ = "low" targ = "high" targ = "high/low"
all	TRUE, FALSE => do/don't show histograms for all data (third row).

Value

... STILL NEEDS A DESCRIPTION ...

Note

The lX vector and/or the value of xmiss can be specified as attributes to the input data matrix before invoking attimp (see cosa2).

See Also

hclust, getclust, and attvalues.

Examples

# acquire object at from the attimp function?
# attvalues(X, group[[2]], at$att[1:5])
# attvalues(X, group[[2]], at$att[6:10], all=T)
# attvalues(X, group[[4]], 44, att.names = 'Att 44', group.name = 'Group 4')

mkampert/rCOSA documentation built on Dec. 23, 2019, 8:21 p.m.