How can I use regression model (MLP or SVM regression or any other) for multi-dimensional output (output matrix has real values and is n x d; d>1)?

hello all,How can I use regression model (MLP or SVM regression or any other) for multi-dimensional output (output matrix has real values and is n x d; d>1). Suppose I have input data matrix (training data) as n x m (m>1) having real values, also output data matrix n x d (d>1) having real values. I want to train my regression model with this data and then predict with unseen data. I want this model should predict all the multidimensional outputs simultaneously.

A suitable example would be more helpful...Thanks

[dpb]

    doc mvregress  % maybe?  (In Statistics Toolbox)