outlier_hierarchy {hierarchicalSets}

Create a new hierarchy based on the outlying elements

Description

This function detects the outlying elements of a HierarchicalSet object and creates a new clustering of the sets only based on these elements. The returned HierarchicalSet object will only contain the outlying elements, thus reducing the universe size. This operation is somewhat similar to principal component analysis, in that the derived clustering is based on the structure not captured by the first clustering, thus modeling the second most dominant feature of the data.

Usage

outlier_hierarchy(set, intersectLimit = 1)

Arguments

set

A HierarchicalSet object

intersectLimit

The proportion of sets an element must be present in to be considered part of the intersect. Standard intersects require it to be present in all sets (intersectLimit = 1), which is also the default

Value

An object of class HierarchicalSet, based on the outliying elements of set

See Also

outlying_elements() for extracting outlying element information from a HierarchicalSet object

Examples

data('twitter')

twitSet <- create_hierarchy(twitter)
twitSetOut <- outlier_hierarchy(twitSet)
twitSetOut