Separate note in a paragraph

Co-authored-by: Adam Schill Collberg <[email protected]>
Co-authored-by:  David Oliver <[email protected]>

Author: 3 people

doc/asciidoc/algorithms/alpha/nodeclassification-pipeline/nodeclassification.adoc

@@ -31,7 +31,9 @@ After configuring the node property steps, one can <<algorithms-ml-nodeclassific
 The training phase (II) proceeds in a manner akin to the standalone version of <<algorithms-ml-nodeclassification,Node Classification>>, where it can train multiple models, select the best one, and report relevant performance metrics.
 
 After <<algorithms-ml-nodeclassification-pipelines-train, training the pipeline>>, a classification pipeline is created.
-This new pipeline inherits the node property steps and feature configuration from the training pipeline and uses them to generate the relevant features for classifying unlabeled nodes. Notice that the classification step can only be done with a classification pipeline (not with a training pipeline).
+This new pipeline inherits the node property steps and feature configuration from the training pipeline and uses them to generate the relevant features for classifying unlabeled nodes.
+
+NOTE: <<algorithms-node-classification-pipelines-predict, Classification> can only be done with a trained classification pipeline (not with a training pipeline).
 
 
 The motivation for using pipelines: