som.nn.run.kernel {som.nn}

calls the specified kernel for som training.

Description

calls the specified kernel for som training.

Usage

som.nn.run.kernel(
  data,
  classes = "no classes",
  kernel = c("internal", "SOM"),
  xdim,
  ydim,
  len = 100,
  alpha = 0.05,
  radius = 1,
  init,
  toroidal = FALSE
)

Arguments

data

numeric matrix or data.frame with training data. Only numeric columns of data.frame are used for training.

classes

character vector with class labels (only necessary for supervised training kernels).

kernel

kernel to be used

xdim

number of neurons in x

ydim

number of neurons in y

len

number of steps to be trained (steps - not epochs!).

alpha

initial learning rate (decreased to 0).

radius

initial radius (decreased to 1).

init

numeric matrix or data.frame with codes for initialisation.

toroidal

true if doughnut-shaped som.

Value

    list with elements \code{codes} and \code{grid}.