We are interested in regression where both input and output vectors are multivariate, in particular linear regression. We know that there is a linear regression function in Weka that only accepts a single value as the output. Although we could train a single-output regressor for each output dimension, this would be very inefficient. We also found a RBFRegressor function that does RBF regression but again with a single output. Is there any function for learning linear or nonlinear regression in Weka that accepts a vector as the output instead of a single value?
We are new to Weka so apologies if we are missing something obvious, but we couldn't find it in the Weka documentation, textbook, or the web.