Running different predictive models in parallel

You an use a Multiply operator to split your training data and from there connect differnet algorithms and have multiple outputs. 

You may also be interested in the "Compare ROCs" operator, which allows you to run mulitple learner types inside and then see how they compare in terms of ROC on your data.  These aren't optimized learners but they will give you a general sense of which types of algorithms look to be better suited to your specific problem.  

You can also combine a optimize with a Select Subprocess operator to always get the better model.

~Martin