Thread Subject: Neural Network- Back Propagation -Matlab function usage

Hi,

I have a task of training a NN ( using back propagation).
I have a set of training data ( files).

Requirement of my system:
The system has 4 inputs ( all fixed i.e constant over a period of time,for 1 data set).
There is one output .( which decreases with time)
The output needs to be fed back.

I want to use the newnarx() function in matlab initially.( as i am not very familiar with theory part)

I went through the examples given in matlab toolbox, but i did not exactly understand how i can give the input data and feedback data to newnarx()
the syntax is

net = newnarx(P,T,ID,OD...)

should P be the vector of my inputs ( 4 values?)

T - target output which is a single value

how can i feed back the output - using delay OD??,

I did check in the matlab doing some trial and error,but if someone can tell me how to use this function or provide some links where i can find simple examples, it will help me in starting off...

Many thanks ..

Also,

How can i provide the initialised weights ( W and V vectors?)

On Mar 26, 3:14=A0am, "Kishore " <kishore3...@yahoo.co.in> wrote:
> Also,
>
> How can i provide the initialised weights ( W and V vectors?)

I don't have newnarx in my old version. However, I don't
see why it should be any different than newff w.r.t
automatically supplying initial weights. After

net =3D newnarx(...);

Just type

net

to see where the weights are stored (e.g., net.IW and
net.LW with newff). If they are empty or zeroed out, just use

net =3D init(net);

(and cross your fingers).

Hope this helps.

Greg

Hi Greg,

Thanks for your reply!!.

Sir, Greg Heath you are right.

I did see the code for newnarx() and it did first use newff() to create the network. And in the end it initialize the weights.
Here is it description
newnarx(PR,[S1 S2...SNl],{TF1 TF2...TFNl},BTF,BLF,PF) takes
PR R x 2 matrix of min and max values for R input elements
ID Input delay vector
OD Output delay vector
Si Size of ith layer, for Nl layers
TFi Transfer function of ith layer (default = 'tansig')
BTF Backpropagation network training function (default = 'trainlm')
BLF Backpropagation weight/bias learning function (default = 'learngdm')
PF Performance function (default = 'mse')

The last line of the newnarx code is as, you suggested
net = init(net);

"Kishore " <kishore3385@yahoo.co.in> wrote in message <gqf9oc$c8v$1@fred.mathworks.com>...
> Hi,
>
> I have a task of training a NN ( using back propagation).
> I have a set of training data ( files).
>
> Requirement of my system:
> The system has 4 inputs ( all fixed i.e constant over a period of time,for 1 data set).
> There is one output .( which decreases with time)
> The output needs to be fed back.
>
> I want to use the newnarx() function in matlab initially.( as i am not very familiar with theory part)
>
> I went through the examples given in matlab toolbox, but i did not exactly understand how i can give the input data and feedback data to newnarx()
> the syntax is
>
> net = newnarx(P,T,ID,OD...)
>
> should P be the vector of my inputs ( 4 values?)
>
> T - target output which is a single value
>
> how can i feed back the output - using delay OD??,
>
> I did check in the matlab doing some trial and error,but if someone can tell me how to use this function or provide some links where i can find simple examples, it will help me in starting off...
>
> Many thanks ..