ComputeMaxInfoGains: Max information gains

In CuCubes: MultiDimensional Feature Selection (MDFS)

Description

Max information gains

Usage

ComputeMaxInfoGains(acceleration.type = "scalar", dimensions = 1,
  divisions = 1, discretizations = 1, seed = 0, range = 1,
  pseudo.count = 0.001, reduce.method = "max", data, decision)

Arguments

acceleration.type	acceleration type ('scalar' for none, 'avx'/'avx2' for use of the AVX/AVX2 instruction set respectively, 'cuda' for CUDA)
dimensions	number of dimensions
divisions	number of divisions
discretizations	number of discretizations
seed	seed for PRNG used during discretizations
range	discretization range (from 0.0 to 1.0)
pseudo.count	pseudo count
reduce.method	discretization reduce method (either "max" or "mean")
data	input data where columns are variables and rows are observations
decision	decision variable as a boolean vector of length equal to number of observations

Value

numeric vector with max information gain for each input variable

Examples

  ComputeMaxInfoGains(data = madelon$data, decision = madelon$decision,
    discretizations = 1, range = 0, divisions = 22, dimensions = 1)

CuCubes documentation built on May 30, 2017, 1:36 a.m.