mmcLoss: Loss function for max-margin clustering

In bmrm: Bundle Methods for Regularized Risk Minimization Package

Description

Loss function for max-margin clustering

Usage

mmcLoss(x, k = 3L, minClusterSize = 1L, groups = matrix(logical(0),
  nrow(x), 0), minGroupOverlap = matrix(integer(0), k, ncol(groups)),
  weight = 1/nrow(x))

Arguments

x	numeric matrix representing the dataset (one sample per row)
k	an integer specifying number of clusters to find
minClusterSize	an integer vector specifying the minimum number of sample per cluster. Given values are reclycled if necessary to have one value per cluster.
groups	a logical matrix for instance grouping (groups[i,j] TRUE when sample i belong to group j).
minGroupOverlap	an integer matrix specifyng the minimum number of instance per cluster for each group.
weight	a weight vector for each instance

Value

the loss function to optimize for max margin clustering of the given dataset

bmrm documentation built on May 2, 2019, 2:49 p.m.