DenseGroupFinder: Dense Group Finder

In aydindemircioglu/Rdrags: High-Dimensional Feature Selection via Dense Relevant Attribute Group Selector

Description

Dense Group Finder

Usage

DenseGroupFinder(data, k = 3, h = 0, useKNN = TRUE, lambda = 0,
  maxIter = 100, eps = 1e-05, kernelType = "gaussian")

Arguments

data	Data to process
k	Neighborhood size, in case useKNN is TRUE.
h	Bandwidth
useKNN	Should k-NN be used to determine the mean distance h?
lambda	Lambda, only used when KernelType is flat
maxIter	Maximum number of iterations
eps	Stopping criterion for convergence.
kernelType	Type of the kernel to be used, either "gaussian" or "flat"

Value

List of groups of (index of) features that are closely related.

aydindemircioglu/Rdrags documentation built on May 14, 2019, 8 a.m.