Prev: projection profile histogram
Next: nonlinear curve fitting with equal and unequal constrains
From: David Cohen on 10 Oct 2009 08:11
Hello all,

I have bulid a Neural Network (RBF) based on data I have gathered using the SVMTrain, I have also test the results from the SVMTrain on other data sets and got satisfactory results.
Now I would like to take the SVMStruct which I got back from the SVMTrain procedure and extract the information from it in order to build an approximation function.
How can it be done?

As my data consists of 4 input parameters I expect to get some function which consists from those parameters with coefficients and weights I am correct?

Any help will be greatly appreciated.
From: Bruno Luong on 10 Oct 2009 11:31
"David Cohen" <dudu.cohen(a)gmail.com> wrote in message <haptkl$k6q$1(a)fred.mathworks.com>...
> Hello all,
>
> I have bulid a Neural Network (RBF) based on data I have gathered using the SVMTrain, I have also test the results from the SVMTrain on other data sets and got satisfactory results.
> Now I would like to take the SVMStruct which I got back from the SVMTrain procedure and extract the information from it in order to build an approximation function.
> How can it be done?
>

Do you mean to know what formula used by svmclassify?

For any x a vector, the class if is the sign of the algebraic distance.

c(x) = sign(d(x)).

The algebraic distance d(x) of x from the hyperplan (in the *feature* space) - scaled by the margin - is:

d(x) = sum K(xi,x)*(alphai.*yi) - 0.5*sum K(xi,s)*(alphai.*yi)

xi is training points; s is (any) support vector
alphai is dual variables
K(.,.) is kernel
x is classify point

Notes
- Matlab classify returns (alphai.*yi) after training.
- The second sum using *one* support vector in the formula can be calculated more accurately by replacing with the mean of the same quantity on *all* support vectors.

But why bother not using svmclassify?

Bruno
From: David Cohen on 10 Oct 2009 12:19
Thank you Bruno,

I am not using svmclassify as this should be implemented in a c++ code in a greater project.
I just use the SVMTrain in order to find out the formula I'll need.

I still don't exactly get what is the formula to extract, Can I use only one support vector? (any one?) what exactly is the K? I have used the default SVMTrain parameters.

[AlphaY, SVs, Bias, Parameters, nSV, nLabel] = SVMTrain(Samples,Labels);
From: Bruno Luong on 10 Oct 2009 14:02
"David Cohen" <dudu.cohen(a)gmail.com> wrote in message <haqc5l$cab$1(a)fred.mathworks.com>...
> Thank you Bruno,
>
> I am not using svmclassify as this should be implemented in a c++ code in a greater project.
> I just use the SVMTrain in order to find out the formula I'll need.
>
> I still don't exactly get what is the formula to extract, Can I use only one support vector? (any one?) what exactly is the K? I have used the default SVMTrain parameters.

K(.,.) is the kernel you are using to train. Look at the doc the see which is used be default. As I said any support vector would do, but it's better to use all of them.

Bruno
 | 
Pages: 1
Prev: projection profile histogram
Next: nonlinear curve fitting with equal and unequal constrains