Hi all:I am trying to fit a SOM which is usually carried out in two steps. At
each step, different parameters (learning rate, neighbourhood and number of
epochs) need to be used.The SOM package is based on SOM-PAK
(http://www.cis.hut.fi/research/papers/som_tr96.ps.Z). In the HTML help
different strategies seem to be correct though I cannot find out how to fine
tune the parameters asociated to each of the two steps:1. In the HELP example
the following appearsdata(yeast)
yeast <- yeast[, -c(1, 11)]
yeast.f <- filtering(yeast)
yeast.f.n <- normalize(yeast.f)
foo <- som(yeast.f.n, xdim=5, ydim=6)
foo <- som(yeast.f.n, xdim=5, ydim=6, topol="hexa", neigh="gaussian")
plot(foo)
As you may notice there are two different lines with foo<-som(...)My question
is as follows: Does this mean that the first line and associated attributes
starting withfoo<-som(..) corresponds exclusively to the 1st phase and the
second line foo<-som(..) corresponds to the second phase and associated
attributes of the same SOM being fitted?2. There are two additional functions
which seem to carry out the same task, namely 'som.init' and 'som.update' The
second question is: Would we reach the same final SOM if the first lineinstead
of foo<-som(..)were foo<-som.init(...)and the second foo<-som(...) were
foo<-som.update(...) ? Could we fine tune the parameters for the fitting
process in this way(first phase with som.init and second phase with
som.update)?Thank youGabriel Ibarra-BerastegiUniversity of the Basque
CountrySPAIN
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
