Hi Arshia,

Support vector machine for regression, in the case you're mentioning, can be implemented as a base classifier of the multioutput algorithm. Class type of the loaded data should belong to the numeric type which can be a challenge for the base algorithm, as most of the platforms that solve multioutput problems deal with nominal or binary type.

HTH.

Samer

I was searching  for multivariate SVM regression - and found the following link

http://svmlight.joachims.org/

It seems that the SVMLight developed at Cornel have developed an SVMLight Struct which is able to do regression with multiple output. But I never tried it out - I will try using it soon. It has a Python and MATLAB wrapper code that facilitates the integration of the code.

I hope it helps

I had not seen that support vector machine to be used for this section but its easy that you learn programing language and development site any algorithm  in line with your needs.

you can get help from Stack Overflow.com to receive code. its a  Community  for question and answer.