newrb [Matlab]

Prev: Finding the position of the maximum of an array ?
Next: how to add two or more signal to scope in simulink

From: aurikel Radzali on 19 Nov 2009 01:36

Greg Heath <heath(a)alumni.brown.edu> wrote in message <9bfd2ece-92b5-432f-a82c-61238af4d20c(a)j4g2000yqe.googlegroups.com>...
> On Nov 16, 4:23 am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > GregHeath<he...(a)alumni.brown.edu> wrote in message <dc9ef436-338f-40fb-aa39-d679e47f6...(a)j11g2000vbi.googlegroups.com>...
> > > On Nov 15, 9:15?pm, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > > > Hi, I am working on radial basis network to predict protein conc. Part of the task is to compare result between newrb & newrbe. I have no problem with the newrb but for newrbe I got this warning:
> >
> > > > Warning: Rank deficient, rank = 5, ?tol = ? 2.7195e-014.
> > > > ?> ?In newrbe>designrbe at 122
> > > > ? ? ?In newrbe at 105
> > > > ? ? ?In RBFnormBnewrbe01_VCC_TRAINOPTIMIZE at 64
> >
> > > > 1.below are my source (newrbe):
> >
> > > > K=TRAINOPTIMIZE;
> > > > [r,c]=size(K);
> >
> > > > GOAL=0.5;%SPREAD=input
> > > > SPREAD=20;
> > > > MN=24; % sample points
> > > > DF=1;
> > > > %net.layers{1}.transferFcn='radbas';
> > > > %net.layers{2}.transferFcn='purelin';
> >
> > > > net=newrbe(P,T,SPREAD);
> > > > Y=sim(net,P);
> >
> > > > 2.below are my source (newrb):
> >
> > > > GOAL=0.5; %SPREAD=input
> > > > SPREAD=20;
> > > > MN=24; % sample points
> > > > DF=1;
> > > > %net.layers{1}.transferFcn='radbas';
> > > > %net.layers{2}.transferFcn='purelin';
> >
> > > > net=newrb(P,T,GOAL,SPREAD);
> > > > Y=sim(net,P);
> >
> > > > note: i got the result as follow:
> > > > NEWRB, neurons = 0, MSE = 0.130203
> >
> > > > where i could be wrong?or did i miss something important?
> >
> > > You say "I have no problem with the newrb".
> > > However, you report "neurons = 0, MSE = 0.130203"
> >
> > > which makes no sense to me. Furthermore, the value 0.13
> > > means nothing unless it is compared to a reference.
> > > What is mean(var(T))?
> >
> > >Greg
> >
> > Hi, actually i have a problem with newrb too except
> > it does not produce warning like the newrbe.
> >
> > --> this is the newrb code:
>
> size(P) = ?
> minmax(P) = ?
> size(T) = ?
> minmax(T) = ?
>
> > GOAL=0.5;
>
> Note: In my version 6.5 this is SSEgoal, NOT MSEgoal;
> -----------> I just put random value, I do not know what value should i choose.

> Why did you choose this value?
>
> For the best constant model y = mean of targets:
>
> y00 = repmat(mean(T,2),1,size(T,2));
> e00 = T-y00;
> SSE00 = sse(e00)
>
> For newrb, it is desirable to have
>
> SSEgoal < SSE00/100
>
> so that
>
> R^2 = 1-SSE/SSE00 > 0.99
> -----------> i do not understand, what is the purpose of this?
>
> > SPREAD=5;
>
> Why did you choose this value?
> ------------------> i just put some value, i don't know what values should i use? (i thought the best spread is obtained through trial and error method)

> How does it compare to
>
> 0.5*mean(median(dist(P,P')))?
> ---------> do you mean compare this and spread=5?
------------->btw, 0.5*mean(median(dist(P,P')))=1.0815
>
> > MN=24; % sample points
> > DF=1;
> >
> > %net.layers{1}.transferFcn='radbas';
> > %net.layers{2}.transferFcn='purelin';
> >
> > net=newrb(P,T,GOAL,SPREAD);
>
> [net, tr] = newrb(P,T,GOAL,SPREAD);
>
-----------------> what is the different between the those two? What is the purpose of putting [net,tr]?

> > Y=sim(net,P);
> > e=T-Y;
> > msetrain=mse(e)
>
> compare with tr.perf(end)
> --------------tr.perf(end)=0.0018

> > --> this is the result:
> > NEWRB, neurons = 0, MSE = 0.130203
>
> If N ~ 24, This is not inconsistent with
> mean(var(T)) = 0.1359.
> ___________> I'm really sorry, but i truly do not understand what do you by If N ~ 24, This is not inconsistent with mean(var(T)) = 0.1359
----------------------->how it suppose to be??

> Where is the tabulation for neurons > 0?
>
> > msetrain =
> >
> > 0.0018
>
> For how many neurons?
> -------------->Number of nodes in hidden layer=2.0000e+000
>
> > why the mse value is different?
>
> Different number of neurons
> What is tr.epoch(end)?
> ---------> 2

> >the mean(var(T)) is 0.1359
>
> MSE(neurons = 0) = (N-1)*mean(var(T'))/N
> -------------> owhh..thats why i get the MSE=0.1302

> Hope this helps.
>
> Greg

==============> Thanks, I still have some questions.

based on your advice, %For newrb, it is desirable to have SSEgoal < SSE00/100
SSE00=3.1249
SSEm=SSE00/100 %produce 0.0312

so i choose my goal 0.0005 % since SSEgoal < SSE00/100. is this right?

From: Greg Heath on 19 Nov 2009 03:24

On Nov 19, 1:36 am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> Greg Heath <he...(a)alumni.brown.edu> wrote in message <9bfd2ece-92b5-432f-a82c-61238af4d...(a)j4g2000yqe.googlegroups.com>...
> > On Nov 16, 4:23 am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > > GregHeath<he...(a)alumni.brown.edu> wrote in message <dc9ef436-338f-40fb-aa39-d679e47f6...(a)j11g2000vbi.googlegroups.com>...
> > > > On Nov 15, 9:15?pm, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > > > > Hi, I am working on radial basis network to predict protein conc.
Part of the task is to compare result between newrb & newrbe. I have
no
problem with the newrb but for newrbe I got this warning:
>
> > > > > Warning: Rank deficient, rank = 5, ?tol = ? 2.7195e-014.
> > > > > ?> ?In newrbe>designrbe at 122
> > > > > ? ? ?In newrbe at 105
> > > > > ? ? ?In RBFnormBnewrbe01_VCC_TRAINOPTIMIZE at 64
>
> > > > > 1.below are my source (newrbe):
>
> > > > > K=TRAINOPTIMIZE;
> > > > > [r,c]=size(K);
>
> > > > > GOAL=0.5;%SPREAD=input
> > > > > SPREAD=20;
> > > > > MN=24; % sample points
> > > > > DF=1;
> > > > > %net.layers{1}.transferFcn='radbas';
> > > > > %net.layers{2}.transferFcn='purelin';
>
> > > > > net=newrbe(P,T,SPREAD);
> > > > > Y=sim(net,P);
>
> > > > > 2.below are my source (newrb):
>
> > > > > GOAL=0.5; %SPREAD=input
> > > > > SPREAD=20;
> > > > > MN=24; % sample points
> > > > > DF=1;
> > > > > %net.layers{1}.transferFcn='radbas';
> > > > > %net.layers{2}.transferFcn='purelin';
>
> > > > > net=newrb(P,T,GOAL,SPREAD);
> > > > > Y=sim(net,P);
>
> > > > > note: i got the result as follow:
> > > > > NEWRB, neurons = 0, MSE = 0.130203
>
> > > > > where i could be wrong?or did i miss something important?
>
> > > > You say "I have no problem with the newrb".
> > > > However, you report "neurons = 0, MSE = 0.130203"
>
> > > > which makes no sense to me. Furthermore, the value 0.13
> > > > means nothing unless it is compared to a reference.
> > > > What is mean(var(T))?
>
> > > >Greg
>
> > > Hi, actually i have a problem with newrb too except
> > > it does not produce warning like the newrbe.
>
> > > --> this is the newrb code:
>
> > size(P) = ?
> > minmax(P) = ?
> > size(T) = ?
> > minmax(T) = ?

Please answer the questions.

> > > GOAL=0.5;
>
> > Note: In my version 6.5 this is SSEgoal, NOT MSEgoal;

> -----------> I just put random value, I do not know what
> value should i choose.

> > Why did you choose this value?
>
> > For the best constant model y = mean of targets:
>
> > y00 = repmat(mean(T,2),1,size(T,2));
> > e00 = T-y00;
> > SSE00 = sse(e00)
>
> > For newrb, it is desirable to have
>
> > SSEgoal < SSE00/100
>
> > so that
>
> > R^2 = 1-SSE/SSE00 > 0.99

> -----------> i do not understand, what is the purpose of this?

R^2 (coefficient of determination) is a measure of the
fraction of output variance that is represented by the model.
See any statistic book that covers regression.

http://en.wikipedia.org/wiki/Coefficient_of_determination

> > > SPREAD=5;
>
> > Why did you choose this value?

> ------------------> i just put some value, i don't know what
values should i use? (i thought the best spread is obtained
> through trial and error method)

True, but why whistle in the dark when you can quickly estimate
a few charcteristic distances to help limit the search?

> > How does it compare to
>
> > 0.5*mean(median(dist(P,P')))?

Typical half-distance between data points. More useful
when the data points are either cluster centers from a previous
cluster analysis or data points selected as neurons.

> ---------> do you mean compare this and spread=5?

Yes.

> ------------->btw, 0.5*mean(median(dist(P,P')))=1.0815

What is it for hidden neurons ?

> > > MN=24; % sample points
> > > DF=1;
>
> > > %net.layers{1}.transferFcn='radbas';
> > > %net.layers{2}.transferFcn='purelin';
>
> > > net=newrb(P,T,GOAL,SPREAD);
>
> > [net, tr] = newrb(P,T,GOAL,SPREAD);
>
> -----------------> what is the different between the those two?
>What is the purpose of putting [net,tr]?

tr is the training record

help newrb
doc newrb

> > > Y=sim(net,P);
> > > e=T-Y;
> > > msetrain=mse(e)
>
> > compare with tr.perf(end)

> --------------tr.perf(end)=0.0018

> > > --> this is the result:
> > > NEWRB, neurons = 0, MSE = 0.130203
>
> > If N ~ 24, This is not inconsistent with
> > mean(var(T)) = 0.1359.

> ___________> I'm really sorry, but i truly do not
understand what do you by If N ~ 24,
This is not inconsistent with mean(var(T)) = 0.1359
>
> ----------------------->how it suppose to be??

If neurons = 0. Then the the output is a constant
equal to the bias. To minimize SSE, the constant must
be mean(T). The resulting MSE is (N-1)*var(T)/N

> > Where is the tabulation for neurons > 0?
>
> > > msetrain =
>
> > > 0.0018
>
> > For how many neurons?
> > -------------->Number of nodes in hidden layer=2.0000e+000
>
> > > why the mse value is different?
>
> > Different number of neurons
> > What is tr.epoch(end)?
> > ---------> 2
> > >the mean(var(T)) is 0.1359
>
> > MSE(neurons = 0) = (N-1)*mean(var(T'))/N
> > -------------> owhh..thats why i get the MSE=0.1302
> > Hope this helps.
>
> > Greg
>
> ==============> Thanks, I still have some questions.
>
> based on your advice, %For newrb, it is desirable to have SSEgoal < SSE00/100
> SSE00=3.1249
> SSEm=SSE00/100 %produce 0.0312
>
> so i choose my goal 0.0005 % since SSEgoal < SSE00/100. is this right?

I would use 0.02.

Hope this helps.

Greg

From: aurikel Radzali on 19 Nov 2009 04:14

Greg Heath <heath(a)alumni.brown.edu> wrote in message <f88c4240-7211-4544-9ca1-f5a6568454cf(a)f20g2000vbl.googlegroups.com>...
> On Nov 19, 1:36 am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > Greg Heath <he...(a)alumni.brown.edu> wrote in message <9bfd2ece-92b5-432f-a82c-61238af4d...(a)j4g2000yqe.googlegroups.com>...
> > > On Nov 16, 4:23 am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > > > GregHeath<he...(a)alumni.brown.edu> wrote in message <dc9ef436-338f-40fb-aa39-d679e47f6...(a)j11g2000vbi.googlegroups.com>...
> > > > > On Nov 15, 9:15?pm, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > > > > > Hi, I am working on radial basis network to predict protein conc.
> Part of the task is to compare result between newrb & newrbe. I have
> no
> problem with the newrb but for newrbe I got this warning:
> >
> > > > > > Warning: Rank deficient, rank = 5, ?tol = ? 2.7195e-014.
> > > > > > ?> ?In newrbe>designrbe at 122
> > > > > > ? ? ?In newrbe at 105
> > > > > > ? ? ?In RBFnormBnewrbe01_VCC_TRAINOPTIMIZE at 64
> >
> > > > > > 1.below are my source (newrbe):
> >
> > > > > > K=TRAINOPTIMIZE;
> > > > > > [r,c]=size(K);
> >
> > > > > > GOAL=0.5;%SPREAD=input
> > > > > > SPREAD=20;
> > > > > > MN=24; % sample points
> > > > > > DF=1;
> > > > > > %net.layers{1}.transferFcn='radbas';
> > > > > > %net.layers{2}.transferFcn='purelin';
> >
> > > > > > net=newrbe(P,T,SPREAD);
> > > > > > Y=sim(net,P);
> >
> > > > > > 2.below are my source (newrb):
> >
> > > > > > GOAL=0.5; %SPREAD=input
> > > > > > SPREAD=20;
> > > > > > MN=24; % sample points
> > > > > > DF=1;
> > > > > > %net.layers{1}.transferFcn='radbas';
> > > > > > %net.layers{2}.transferFcn='purelin';
> >
> > > > > > net=newrb(P,T,GOAL,SPREAD);
> > > > > > Y=sim(net,P);
> >
> > > > > > note: i got the result as follow:
> > > > > > NEWRB, neurons = 0, MSE = 0.130203
> >
> > > > > > where i could be wrong?or did i miss something important?
> >
> > > > > You say "I have no problem with the newrb".
> > > > > However, you report "neurons = 0, MSE = 0.130203"
> >
> > > > > which makes no sense to me. Furthermore, the value 0.13
> > > > > means nothing unless it is compared to a reference.
> > > > > What is mean(var(T))?
> >
> > > > >Greg
> >
> > > > Hi, actually i have a problem with newrb too except
> > > > it does not produce warning like the newrbe.
> >
> > > > --> this is the newrb code:
> >
> > > size(P) = ?
> > > minmax(P) = ?
> > > size(T) = ?
> > > minmax(T) = ?
>
>
> Please answer the questions.
---->size(P) =3 24
size(T) =1 24

>
> > > > GOAL=0.5;
> >
> > > Note: In my version 6.5 this is SSEgoal, NOT MSEgoal;
>
> > -----------> I just put random value, I do not know what
> > value should i choose.
>
> > > Why did you choose this value?
> >
> > > For the best constant model y = mean of targets:
> >
> > > y00 = repmat(mean(T,2),1,size(T,2));
> > > e00 = T-y00;
> > > SSE00 = sse(e00)
> >
> > > For newrb, it is desirable to have
> >
> > > SSEgoal < SSE00/100
> >
> > > so that
> >
> > > R^2 = 1-SSE/SSE00 > 0.99
>
> > -----------> i do not understand, what is the purpose of this?
>
> R^2 (coefficient of determination) is a measure of the
> fraction of output variance that is represented by the model.
> See any statistic book that covers regression.
>
> http://en.wikipedia.org/wiki/Coefficient_of_determination
>
> > > > SPREAD=5;
> >
> > > Why did you choose this value?
>
> > ------------------> i just put some value, i don't know what
> values should i use? (i thought the best spread is obtained
> > through trial and error method)
>
> True, but why whistle in the dark when you can quickly estimate
> a few charcteristic distances to help limit the search?
>
------------------>how do i estimate characteristic distances to help limit the search? i even do not know what no should i start with.

> > > How does it compare to
> >
> > > 0.5*mean(median(dist(P,P')))?
>
> Typical half-distance between data points. More useful
> when the data points are either cluster centers from a previous
> cluster analysis or data points selected as neurons.
>
> > ---------> do you mean compare this and spread=5?
>
> Yes.
>
> > ------------->btw, 0.5*mean(median(dist(P,P')))=1.0815

-----------> is this some kind of guideline to choose the spread? the spread should be around that value, it is?less or more but not too far away?

> What is it for hidden neurons ?
>
> > > > MN=24; % sample points
> > > > DF=1;
> >
> > > > %net.layers{1}.transferFcn='radbas';
> > > > %net.layers{2}.transferFcn='purelin';
> >
> > > > net=newrb(P,T,GOAL,SPREAD);
> >
> > > [net, tr] = newrb(P,T,GOAL,SPREAD);
> >
> > -----------------> what is the different between the those two?
> >What is the purpose of putting [net,tr]?
>
> tr is the training record
>
> help newrb
> doc newrb
>
> > > > Y=sim(net,P);
> > > > e=T-Y;
> > > > msetrain=mse(e)
> >
> > > compare with tr.perf(end)
>
> > --------------tr.perf(end)=0.0018
>
> > > > --> this is the result:
> > > > NEWRB, neurons = 0, MSE = 0.130203
> >
> > > If N ~ 24, This is not inconsistent with
> > > mean(var(T)) = 0.1359.
>
> > ___________> I'm really sorry, but i truly do not
> understand what do you by If N ~ 24,
> This is not inconsistent with mean(var(T)) = 0.1359
> >
> > ----------------------->how it suppose to be??
>
> If neurons = 0. Then the the output is a constant
> equal to the bias. To minimize SSE, the constant must
> be mean(T). The resulting MSE is (N-1)*var(T)/N
>
---------->forgive me, i still don't understand.
mean(T)=0.5389
mean(var(T)) = 0.1359
MSE i= (N-1)*var(T)/N= 0.1302 so?

--->To minimize SSE, 'the constant' must be mean(T). 'the constant' refer to?

> > > Where is the tabulation for neurons > 0?
> >
> > > > msetrain =
> >
> > > > 0.0018
> >
> > > For how many neurons?
> > > -------------->Number of nodes in hidden layer=2.0000e+000
> >
> > > > why the mse value is different?
> >
> > > Different number of neurons
> > > What is tr.epoch(end)?
> > > ---------> 2
> > > >the mean(var(T)) is 0.1359
> >
> > > MSE(neurons = 0) = (N-1)*mean(var(T'))/N
> > > -------------> owhh..thats why i get the MSE=0.1302
> > > Hope this helps.
> >
> > > Greg
> >
> > ==============> Thanks, I still have some questions.
> >
> > based on your advice, %For newrb, it is desirable to have SSEgoal < SSE00/100
> > SSE00=3.1249
> > SSEm=SSE00/100 %produce 0.0312
> >
> > so i choose my goal 0.0005 % since SSEgoal < SSE00/100. is this right?
>
> I would use 0.02.
>
> Hope this helps.
>
> Greg

From: Greg Heath on 19 Nov 2009 07:01

On Nov 19, 4:14 am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> Greg Heath <he...(a)alumni.brown.edu> wrote > > > > size(P) = ?
> > > > minmax(P) = ?
> > > > size(T) = ?
> > > > minmax(T) = ?
>
> > Please answer the questions.
>
> ---->size(P) =3 24
> size(T) =1 24

MINMAX??

Greg

From: aurikel Radzali on 19 Nov 2009 07:44

Greg Heath <heath(a)alumni.brown.edu> wrote in message <85397222-7dfd-4f31-a3f7-6e2a088e47e9(a)p33g2000vbn.googlegroups.com>...
> On Nov 19, 4:14?am, "aurikel Radzali" <suriar...(a)gmail.com> wrote:
> > Greg Heath <he...(a)alumni.brown.edu> wrote > > > > size(P) = ?
> > > > > minmax(P) = ?
> > > > > size(T) = ?
> > > > > minmax(T) = ?
> >
> > > Please answer the questions.
> >
> > ---->size(P) =3 ? ?24
> > ? ? ? ?size(T) =1 ?24
>
> MINMAX??
>
> Greg

MINMAXP =

0 1
0 1
0 1

MINMAXT =

0 1

| Next | Last
Pages: 1 2
Prev: Finding the position of the maximum of an array ?
Next: how to add two or more signal to scope in simulink