At last, the “data preparation” meta-node prepares the text data for classification, by converting all
description text strings into Document type, extracting the bag of words, getting rid of all noncontent
related words, reducing words to their stems, deleting conflicting tags, and finally extracting
the keywords for each document (Fig. 18). The Integer Input Quickform node defines the number of
folds for the cross-validation node hosting the classification algorithm down the workflow.