[R] Softmax in nnet

[G Ilhamto]

(this is a resend)

Hi R help,

 I run my data in nnet with skip layer, factor response (with 0 & 1
 values) and explicitly put softmax=T to compare the result of the
 default nnet with no softmax specification. I assume this should give
 me the same result. I got the result from the default one, but not the
 softmax version and I got the error message that I did not quite
 understand.

 test6.nn.skipT.softm.Yfac <- nnet(Yfac~ X1 +.. +X8, skip=T, size=0,
 softmax=T, data = train.set)

 Error in nnet.default(x, y, w, entropy = TRUE, ...) : no weights to fit

 In addition: Warning messages:
 1: In if (softmax) { :  the condition has length > 1 and only the
 first element will be used
 2: In if (skip) net <- add.net(net, seq(1, net$n[1]), seq(1 + net$n[1] +  :
 the condition has length > 1 and only the first element will be used

 When I specify the weights (0.1 just for a try)
 test6.nn.skipT.softm.Yfac <- nnet(Yfac~ X1 +.. +X8, skip=T, size=0,
 weights= 0.1, softmax=T, data = train.set)

 I got another error message:
 Error in model.frame.default(formula = Yfac ~ HusYEduc + AgeRespd +
 muslimat +  :
 variable lengths differ (found for '(weights)')
 -----------------

 Q: Does softmax apply to two-category response?
    What is softmax require?

 Thank you,
 Ilham

On 2/18/08, G Ilhamto <gilhamto at gmail.com> wrote:
> Hi R help,
>
>  I run my data in nnet with skip layer, factor response (with 0 & 1
>  values) and explicitly put softmax=T to compare the result of the
>  default nnet with no softmax specification. I assume this should give
>  me the same result. I got the result the default one, but not the
>  softmax version and I got the error message that I did not quite
>  understand.
>
>  test6.nn.skipT.softm.Yfac <- nnet(Yfac~ X1 +.. +X8, skip=T, size=0,
>  softmax=T, data = train.set)
>
>  Error in nnet.default(x, y, w, entropy = TRUE, ...) : no weights to fit
>
>  In addition: Warning messages:
>  1: In if (softmax) { :  the condition has length > 1 and only the
>  first element will be used
>  2: In if (skip) net <- add.net(net, seq(1, net$n[1]), seq(1 + net$n[1] +  :
>   the condition has length > 1 and only the first element will be used
>
>  When I specify the weights (0.1 just for a try)
>  test6.nn.skipT.softm.Yfac <- nnet(Yfac~ X1 +.. +X8, skip=T, size=0,
>  weights= 0.1, softmax=T, data = train.set)
>
>  I got another error message:
>  Error in model.frame.default(formula = Yfac ~ HusYEduc + AgeRespd +
>  muslimat +  :
>  variable lengths differ (found for '(weights)')
>  -----------------
>
>  Q: Does softmax apply to two-category response?
>      What is softmax require?
>
>  Thank you,
>  Ilham
>