I installed WEKA and downloaded wlsvm, added the jars, libsvm.jar and wlsvm.jar to CLASSPATH thus, succesfully integrated LibSVM into Weka Environment. I generated random numeric attributes (features) and categorical -of course- classes and tried run LibSVM but I got this error "problem evaluating classifier: rand". I have no idea why I got this error, I looked it up on the internet and didn't end up with any solution. would you please help me? By the way, I am so new to WEKA, Java and LibSVM.
LIBSVM is a great library. However, in my experience the best, most up to date and most tested version of LIBSVM is the normal, C++ version of LIBSVM. I would suggest you try downloading and testing the random data you generated in the normal LIBSVM. This doesn't even involve C++ coding all you have to do is generate a training file and a testing file in the LIBSVM input format.
Doing so will allow you to more effectively troubleshoot what is going on. If it doesn't work with the C++ version, then we can definitely check that, if it does work with the C++ version we can be sure that the problem is the (1) interaction with WEKA, (2) an old bug in LIBSVM java that has not been fixed, (3) what you call wlsvm.jar (which I'm not sure what it is) or a mixture of these things.