GroupPostHocs.dist: Groups elements of the given 'dissimilarity matrix'

In adamryczkowski/PostHocGroups: Approximation of paired post-hoc results as mutualy exclusive sets

Description

This is the most general way to invoke the algorithm. User needs to provide distance matrix that is interpreted as dissimilarity matrix, with value that is interpreted as a measure of dissimilarity between items:

matrix value >0

Two items are similar

matrix value <0

Two items are different

Usage

## S3 method for class 'dist'
GroupPostHocs(dissimiliarity.matrix, means = NULL,
  max.recursive.level = 1, solutions.count = 1)

Arguments

dissimiliarity.matrix	Object of dist class.
means	Optional parameter. Requires numeric vector of the size of number of columns of the dissimiliarity.matrix or NULL (default). If specified, function will sort the dissimiliarity.matrix according to those values during the pre-process stage of the algorithm.
max.recursive.level	Defaults to 1 for flat representation. If more, algorithm will try to recursively subpartition groups that have more than 2 elements.
solutions.count	Number of sultions to report. Defaults to 1, i.e. only the best solution will be printed.

Details

To get meaningfull results, user should either provide matrix with items (columns/rows) of the matrix are sorted according to the corresponding value of the tested item, or provide the means argument and the algorithm will sort the elements internally.

Value

Object of class PostHocGroups. If only one possible solution was requested, the object has also class PostHocGroup.

adamryczkowski/PostHocGroups documentation built on May 10, 2019, 5:50 a.m.