Illustrative diagram of data flow through the pipeline.

<p>Data is initially partitioned into discovery and classification sets. The classification set is further partitioned into training and validation sets. After WordSpy elicits motifs using the discovery set, fuzznuc or fuzzpro counts corresponding motif occurrences in the remaining data. The training data counts are used to train a classifier, while the validation data counts are used to determine performance (e.g. AUC) of the learned classifier.