Hi @Laurel,
*the learner created a prediction model (so the logic of the decision tree)
For the predictor node you now need to input the data you want to predict (i guess the upper port of the Partitioning Node)
*the Random Forest Predictor needs and input model (from an Random Forest Lerner Node)
and the Random Forest Predictor (Regression) needs an input model from the Random Forest Lerner (Regression) Node
You should check the example workflows:
You should also check the KNIME Hub
or Nodepit
to look into the configuration description and find already created workflows by the community (so you can check out how they build there models/configured the nodes)
In general - if nodes have an input port, then you need to fill that connection with some input (some exceptions exists - but then they are markt as optional in the node description)