Put the Extreme Learning Machine algorithm into the WEKA
The algorithm need three parameters:
active function, like "sig" "sin".
number of hidden neurons, default 20, more bigger more better but slowly.
random seed, ensure the algorithm can get the result again.
You can classifier the Nominal and Numeric class attribute.
the algorithm is good, it is worthy to try it .
need the matrix jar package.
mtj-0.9.13.jar netlib-java-0.9.3.jar blas-0.8.jar arpack-combo-0.1.jar
and the weka jar package
weka.jar weka-src.jar