SimpleMKL.classification: Simple MKL

In RMKL: Multiple Kernel Learning for Classification or Regression Problems

Description

This function conducts Simple MKL for precomputed gramm matrices

Usage

SimpleMKL.classification(k, outcome, penalty, tol = 10^(-4),
  max.iters = 1000)

Arguments

k	list of Gramm matrices
outcome	vector of binary outcome -1 and 1
penalty	ppenalty of the smoothness of the resulting desicion rules
tol	change between to iterations is smaller than this, algorithms is considered to have converged
max.iters	maximum number of allowed iteratons

Value

gamma weight vector for the importnace of each kernel

alpha coeffiencents of the dual of MKL

time total amount of time to train model

max.iters Numvber of iterations to reach convergence criteria

Examples

library(kernlab)
library(caret)
library(RMKL)
#Load data
data(benchmark.data)
example.data=benchmark.data[[1]]
# Split samples into training and test sets 
training.samples=sample(1:dim(example.data)[1],floor(0.7*dim(example.data)[1]),replace=FALSE)
# Set up cost parameters and kernels 
C=100
kernels=rep('radial',3)
degree=rep(0,3)
scale=rep(0,3)
sigma=c(0,2^seq(-3:0))
K=kernels.gen(example.data[,1:2], training.samples, kernels, degree, scale, sigma)
K.train=K$K.train
SimpleMKL.classification(K.train,example.data[training.samples,3], C)

Example output

Loading required package: lattice
Loading required package: ggplot2

Attaching package: ‘ggplot2’

The following object is masked from ‘package:kernlab’:

    alpha

$gamma
[1] 0 1 0

$iters
[1] 2

$alpha
  [1] 2.145553e-03 4.408910e-05 3.305602e+01 9.999992e+01 1.510327e-04
  [6] 1.006381e+01 5.955634e+01 1.367328e-04 7.490723e-05 9.357457e+01
 [11] 3.009480e-05 9.999999e+01 3.355234e+01 4.510966e-05 4.157061e+01
 [16] 4.322098e-05 2.878140e+01 3.517140e+01 5.006819e-05 8.655379e+00
 [21] 7.704790e+01 3.096216e-03 4.199576e-04 3.663009e-04 9.485126e+01
 [26] 2.022167e-01 1.063272e-05 3.830590e+01 5.403788e+01 4.769932e-05
 [31] 4.550526e+01 2.488669e+01 4.254738e-05 9.330425e-05 1.000000e+02
 [36] 1.938196e+01 7.059709e-06 2.879869e-04 1.000000e+02 1.881237e+01
 [41] 9.999999e+01 7.290926e+01 1.368251e-05 1.833999e-03 2.513803e+00
 [46] 1.215574e+01 6.457470e+01 1.224067e-04 1.723929e+01 1.491627e+01
 [51] 1.192191e-05 4.693348e+01 8.235141e-05 9.999999e+01 9.195460e+00
 [56] 2.384703e+01 1.000000e+02 3.568320e+01 3.981720e-05 6.119435e+01
 [61] 1.099335e-04 9.335132e+01 3.283527e+01 1.831674e-05 5.352280e-05
 [66] 3.922601e+01 2.661056e+01 9.999997e+01 1.136354e-05 1.297337e-02
 [71] 3.788881e-05 2.106342e-05 1.745993e+01 8.288365e+01 1.675527e-05
 [76] 8.198266e+01 8.287029e+01 1.229353e+01 5.506263e+01 2.461231e+01
 [81] 5.173841e+00 2.725858e+01 8.698796e+01 2.131988e+01 1.363580e+01
 [86] 3.905219e-05 3.075299e+01 1.188970e-05 7.514731e-04 1.984472e+01
 [91] 3.463592e-05 1.579760e+01 9.631525e-06 1.191972e-04 1.361960e-04
 [96] 1.528748e-05 7.926626e+01 3.415773e-04 8.772297e-04 1.000000e+02
[101] 3.373321e+01 1.232445e-04 9.999999e+01 9.306351e+01 4.843026e+01
[106] 3.344064e+01 6.260597e+01 4.263392e-05 3.717509e+01 1.000000e+02
[111] 1.962077e+01 7.166229e+01 6.048994e+00 2.966697e-05 3.720834e-05
[116] 3.955233e-04 2.736186e+01 1.848997e-05 3.661057e-05 6.439493e+01
[121] 2.902792e+01 8.906003e-06 9.999999e+01 3.642794e+01 1.666375e-04
[126] 7.206920e+01 4.882937e+01 1.023735e+01 2.965309e+01 3.079292e+01
[131] 1.740635e-05 2.116364e-04 9.999998e+01 3.207111e+01 2.579153e+01
[136] 8.814107e-06 6.054023e-05 1.096759e-04 8.122470e+01 4.045217e+01

$b
[1] 0.03714253

$gamma_all
$gamma_all[[1]]
[1] 0.3333333 0.3333333 0.3333333

$gamma_all[[2]]
[1] 0.0000000 0.5394908 0.4605092

$gamma_all[[3]]
[1] 0 1 0

RMKL documentation built on May 2, 2019, 7:55 a.m.