subspace (version 1.0.4)

plot.subspace_clustering: Plotting Subspace Clusterings

Description

Plotting for Subspace clusterings as generated by the package subspace.

Generates a 2d-scatterplot with interactive controls to select the dimensions that should be plotted. This visualization is created using the ggvis package and is therefore also compatible with shiny.

Usage

"plot"(x, data, color_by = "mix", standardcolors = c("#1F77B4", "#FF7F0E", "#2CA02C", "#D62728", "#9467BD", "#8C564B", "#E377C2", "#7F7F7F", "#BCBD22", "#17BECF", "#000000"), tooltip_on = "hover", ...)

Arguments

x

an S3-Object of type subspace_clustering as generated by any of the functions of the subspace package

data

The original data matrix on which the clustering was performed.

color_by

a parameter indicating how a point that is in multiple clusters should be colored. If "mix" is selected, the point will be colored as a mixture of the colors of both of the clusters that the point is in. If "any" is selected, a random color is selected from the colors of all the clusters that the point is in.

standardcolors

a vector of strings representing HTML-Colors that will be used to color the points by cluster assignment. Noise will be colored with the last color in the vector.

tooltip_on

decides if tooltips should be shown on "hover" or on "click"

...

this is passed on to ggvis::layer_points and can be used to change, for example the size of the points

Value

a ggvis object. If the return value is not used, a plot will be shown, but the returned plot can also be altered using ggvis

Examples

#Load the example dataset for this package
data("subspace_dataset")
#Load the true clustering for this dataset
path_to_clustering <- paste(path.package("subspace"),"/extdata/subspace_dataset.true",sep="")
clustering <- clustering_from_file(file_path=path_to_clustering)
#also generate a clustering with one of the algorithms
clustering2 <- CLIQUE(subspace_dataset,tau=0.2)

#now plot the generated clustering
plot(clustering2,subspace_dataset)
#plot the true clustering with small points
plot(clustering,subspace_dataset,size=0.1)

#Now plot the points with a different shape.
#This requires the workaround that was discussed in "Notes"
p <- ggvis::prop(property="shape",x="cross")
plot(clustering,subspace_dataset,props=p)