MATLAB Central Newsreader - help me to train the following in neuralnetwork

help me to train the following in neuralnetwork

Steve — Mon, 12 Nov 2007 11:15:55 -0500

Hello all,

I tried to train the foll. network i got an error,could
anybodysuggest me,

[P]=[p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14 p15 ];

[T]=[t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 t11 t12 t13 t14 t15 ];

[NVarin Ncases]=size(P)
[NVarin Ncases]=size(T)

net = newff(minmax(P),[1 0 1],
{'tansig' 'purelin','traingdm'});
Y = sim(net,P);
plot(P,T,P,Y,'o')
hold on
net.trainParam.epochs = 100;
net.trainParam.goal = 1e-6;
net = train(net,P,T);
Y = sim(net,P);
plot(P,T,P,Y,'o')

Where pi= 1 x 20
corresponding target ti= 1000 x 2 (i= 1 to 15)

Thankyou.

Re: help me to train the following in neuralnetwork

David — Mon, 12 Nov 2007 12:26:27 -0500

"Steve " <senthilkumaar_s@yahoo.de> wrote in message
<fh9clb$ivr$1@fred.mathworks.com>...
>
> Hello all,
>
> I tried to train the foll. network i got an error,could
> anybodysuggest me,
>
>
> [P]=[p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14
p15 ];
>
> [T]=[t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 t11 t12 t13 t14
t15 ];
>
> [NVarin Ncases]=size(P)
> [NVarin Ncases]=size(T)
>
> net = newff(minmax(P),[1 0 1],
> {'tansig' 'purelin','traingdm'});
> Y = sim(net,P);
> plot(P,T,P,Y,'o')
> hold on
> net.trainParam.epochs = 100;
> net.trainParam.goal = 1e-6;
> net = train(net,P,T);
> Y = sim(net,P);
> plot(P,T,P,Y,'o')
>
> Where pi= 1 x 20
> corresponding target ti= 1000 x 2 (i= 1 to 15)
>
> Thankyou.

i would suggest that you might want to say what the error
is you got and where.

Re: help me to train the following in neuralnetwork

Steve — Mon, 12 Nov 2007 12:38:36 -0500

The error is in minmax

Re: help me to train the following in neuralnetwork

Steve — Mon, 12 Nov 2007 12:38:43 -0500

The error is in minmax

Re: help me to train the following in neuralnetwork

David — Mon, 12 Nov 2007 13:27:50 -0500

"Steve " <senthilkumaar_s@yahoo.de> wrote in message
<fh9hgj$jj0$1@fred.mathworks.com>...
> The error is in minmax

and just what might the error be?

Re: help me to train the following in neuralnetwork

Steve — Mon, 12 Nov 2007 13:40:56 -0500

I got where the eroor is, but i dot how to overcome now,

The error is in pi matrix , I have both numerical and string
for eg. p1=[1 2 3 4 5 6 7 8 'yes' 10 11 'yes' 3 4 5 ]

how to overcome this ??

Re: help me to train the following in neuralnetwork

David — Mon, 12 Nov 2007 14:05:18 -0500

"Steve " <senthilkumaar_s@yahoo.de> wrote in message
<fh9l58$bui$1@fred.mathworks.com>...
> I got where the eroor is, but i dot how to overcome now,
>
> The error is in pi matrix , I have both numerical and
string
> for eg. p1=[1 2 3 4 5 6 7 8 'yes' 10 11 'yes' 3 4 5 ]
>
> how to overcome this ??

you can overcome it by writing your own function that can
decide which is bigger, 'yes' or 11. and which is
smaller, 'yes' or 1.

Re: help me to train the following in neuralnetwork

Steve — Mon, 12 Nov 2007 16:06:27 -0500

Hi ,

I did as u suggested and now i got some other error,

could u please suggest me how to give the t value in newff
function , bcoz i have two columns as output,

my [T]=[t1 t2....t15]
Size of T matrix is 700 x 30

thank you.

Re: help me to train the following in neuralnetwork

David — Mon, 12 Nov 2007 16:19:52 -0500

"Steve " <johnsonsteve12@yahoo.com> wrote in message
<fh9tm3$r2$1@fred.mathworks.com>...
> Hi ,
>
> I did as u suggested and now i got some other error,
>
> could u please suggest me how to give the t value in
newff
> function , bcoz i have two columns as output,
>
> my [T]=[t1 t2....t15]
> Size of T matrix is 700 x 30
>
> thank you.

exactly what is the error you are getting now?

Re: help me to train the following in neuralnetwork

Steve — Mon, 12 Nov 2007 16:44:06 -0500

Error is:

Input range has 2column matrix

Re: help me to train the following in neuralnetwork

Greg Heath — Wed, 14 Nov 2007 13:14:23 -0500

On Nov 12, 6:15 am, "Steve " <senthilkumaa...@yahoo.de> wrote:
> Hello all,
>
> I tried to train the foll. network i got an error,could
> anybodysuggest me,
>
> [P]=[p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14 p15 ];
>
> [T]=[t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 t11 t12 t13 t14 t15 ];
>
> [NVarin Ncases]=size(P)
> [NVarin Ncases]=size(T)

[NVarout Ncases]=size(T)

Nvarin = Nvarout = 1
Ncases = 15

> net = newff(minmax(P),[1 0 1],
> {'tansig' 'purelin','traingdm'});

[1 0 1] is an error it indicates you have

1 node in the 1st hidden layer
0 nodes in the second hidden layer!
1 output node

What configuration do you really want?

I recommend using one hidden layer via [H 1]
and finding a suitable value for H by trial and error.

> Y = sim(net,P);
> plot(P,T,P,Y,'o')
> hold on
> net.trainParam.epochs = 100;
> net.trainParam.goal = 1e-6;

Why 1e-6? Why not some small fraction of var(T)?

> net = train(net,P,T);
> Y = sim(net,P);
> plot(P,T,P,Y,'o')
>
> Where pi= 1 x 20
> corresponding target ti= 1000 x 2 (i= 1 to 15)

This makes no sense to me.

What are Nvarin, Nvarout and Ncases?

Hope this helps.

Greg

P.S. Searching on

greg-heath pretraining advice

may help