A bunch of ML regression models are defined only for predicting the value of a single variable. Or have standard implementation that are only for the univariate case. For example support vector machine and random forest regression models.
I am contemplating the naive (but generic) way to extend them for multivariate regression, via training seperate models for each output variables.
Am I right in saying this would only be correct if the output variables are conditionally independent, when conditioning upon the input feature variables?