featurespace: Plot method for 'trainOcc' object

In benmack/oneClass: One-class classification in the absence of test data

Description

Plot the decision values of a trainOcc object in its two dimensional feature space.

Usage

featurespace(x, thresholds = NULL, x.tr = NULL, positive = NULL,
  borders = NULL, nCells = c(100, 100), main = NULL, column = NULL,
  expandColors = TRUE, ...)

Arguments

x	An object of class trainOcc.
thresholds	a scalar or numeric vector used to plot additional contour lines.
x.tr	data to be plotted. must have a column y with labels.
positive	the class of interest in x.tr$y
borders	names list (x, y) with two vectors of length two specifiying the borders of the feature space in horizontal/x- and vertical/y- directions.
nCells	number of cells in horizontal/x- and vertical/y- directions.
main	character string for the plot title
column	select a column in case the model x is not a trainOcc object and returns more then one column, default is 1.
expandColors	if the predictive values are not positive and negative numbers (as in case of the ocsvm and biasedsvm) set this argument to FALSE
...	other arguments that can be passed to predict.

Details

featurespace

This function is an adaptation of plot.ksvm from the kernlab package.

Value

a plot of the feature space

Examples

## Not run: 
data(bananas)
tr <- bananas$tr
### underfitted model
oc <- trainOcc ( x = tr[, -1], y = tr[, 1], 
                 tuneGrid=expand.grid(sigma=0.1, 
                                      cNeg=.1, 
                                      cMultiplier=100))
featurespace(oc, th=0) 

### overfitted model
oc <- trainOcc ( x = tr[, -1], y = tr[, 1], 
                 tuneGrid=expand.grid(sigma=10, 
                                      cNeg=.1, 
                                      cMultiplier=100))
featurespace(oc, th=0) 

## End(Not run)

benmack/oneClass documentation built on Dec. 15, 2020, 7:38 p.m.