Prev: rotation of a camera
Next: xpc, tcp/ip connect error
From: Thomas on 5 Mar 2010 09:09
Hi!

I have a question I could not sort out by the matlab documentation.

I train a neural network, a feedforward MLP, and it seems that all the learning algorithms such as the backpopagation as well as the scaled conjugate gradient algorithm apply batch learning (updating of the weights are done after all the patterns has been processed through the network).

Does anyone know how to change this to online training?
I have pretty big data sets and the batch learning algorithm takes some really long time.

I really appreciate any help!

Thanks
Thomas
From: Greg Heath on 5 Mar 2010 19:22
On Mar 5, 9:09 am, "Thomas " <thomas.pa...(a)uis.no> wrote:
> Hi!
>
> I have a question I could not sort out by the matlab documentation.
>
> I train a neural network, a feedforward MLP, and it seems that all the learning algorithms such as the backpopagation as well as the scaled conjugate gradient algorithm apply batch learning (updating of the weights are done after all the patterns has been processed through the network).
>
> Does anyone know how to change this to online training?
> I have pretty big data sets and the batch learning algorithm takes some really long time.
>
> I really appreciate any help!

Using MATLAB 6.5:

help nnet
....
Training functions.
trainb - Batch training with weight & bias learning rules.
trainbfg - BFGS quasi-Newton backpropagation.
trainbr - Bayesian regularization.
trainc - Cyclical order incremental training w/learning
functions.
traincgb - Powell-Beale conjugate gradient backpropagation.
traincgf - Fletcher-Powell conjugate gradient backpropagation.
traincgp - Polak-Ribiere conjugate gradient backpropagation.
traingd - Gradient descent backpropagation.
traingdm - Gradient descent with momentum backpropagation.
traingda - Gradient descent with adaptive lr backpropagation.
traingdx - Gradient descent w/momentum & adaptive lr
backpropagation.
trainlm - Levenberg-Marquardt backpropagation.
trainoss - One step secant backpropagation.
trainr - Random order incremental training w/learning functions.
trainrp - Resilient backpropagation (Rprop).
trains - Sequential order incremental training w/learning
functions.
trainscg - Scaled conjugate gradient backpropagation.

So, ... take a look at

doc trainc
help trainc

doc trainoss
help trainoss

doc trains
help trains

Hope this helps.

Greg

 | 
Pages: 1
Prev: rotation of a camera
Next: xpc, tcp/ip connect error