[R] neural package

From: Jeszenszky Peter <jeszenszky.peter_at_inf.unideb.hu>
Date: Wed, 05 Nov 2008 00:08:20 +0100

   Hello,
   I am very happy to see that the Rprop neural network training    algorithm is now available in the neuralnet package.    I have a few questions about the package and a few feature    requests also.
   According to the documentation the startweights argument    of the neuralnet function is

           "a vector containing starting values for the weights.
           The weights will not be randomly initialized."
   If the weights of the network are not initialized randomly, do    I have to initialize them explicitly using this argument ?    If the answer is yes, how to initialize them? (None of the    examples in the documentation uses the startweights argument.)    The err.fct argument of the neuralnet function may have the    special "ce" value, that stands for cross entropy minimization.    I think that it is useful if the output layer consists of softmax    neurons. Are softmax output units supported?    My feature requests are the following. (I think that it would    be easy to implement them. Actually, I have done it myself.)    The plot.nn function accepts rep="best". It would be convenient    to pass par="best" to the function compute also that currently    accepts only an integer as the value of the rep argument.    Finally, it would be nice to store the number of the best    repetition found as a component in the neuralnet object (for    examlpe, as $rep.best). This can be accessed only as

           rep.best <- which.min(net$result.matrix["error",])    currently, which is quite technical.
   Best regards,
   Peter Jeszenszky



R-help_at_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. Received on Tue 04 Nov 2008 - 23:12:03 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Tue 04 Nov 2008 - 23:30:23 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive