ModelSV {svcR}R Documentation

Computation of lagrange coefficients and radius of the model

Description

Computes kernel matrix, lagrange coefficients, support vectors and radius

Usage

## S4 method for signature 'numeric':
ModelSV.compute(x ,
          MatriceKernel = NULL,
          MatriceK      = NULL,
          Nu            = 1,
          nlin          = 1,
          MaxIter       = 2,
          MaxValA       = 2,
          AroundNull    = 0.01,
          AroundNullVA  = 0.01  )

## S4 method for signature 'matrix':
OptimQuadProgWcluster(MatriceKernel ,
          Nu            = 1,
          MaxValA       = 2,
          MinW          = 0.0001 )

Arguments

x method of lagrange computation 1 or 2
MatriceKernel kernel matrix with vector format
MatriceK kernel matrix with vector format
Nu svc parameter
nlin number
MaxIter maximum iteration for coefficients computation
MaxValA number of neigbours on the grid
AroundNull almost null parameter
AroundNullVA almost null parameter for coefficients
MinW min value for coefficients

Value

An S4 object of class ModelSV The object is the svc model along with the slots :

VectorWA lagrange coefficients : VectorsYA$A
RadiusC radius of the hypersphere containing data
SmallR residu of radius

OptimQuadProgWcluster return a list containing lagrange coefficients.
slots can be accessed by object@slot

Author(s)

Nicolas Turenne - INRA France nicolas.turenne@jouy.inra.fr

References

N.Turenne , Some Heuristics to speed-up Support Vector Clustering , technical report 2006, INRA, France http://migale.jouy.inra.fr/~turenne/svc.pdf

Examples


## exemple with iris data

# model computing 
fmc = findModelCluster.Test(); 

fmc@WVectorsYA$A; # lagrange coefficients


[Package svcR version 1.4.2 Index]