Features vs. bag of words

  1. In qiime2's multinomial Naive Bayes classifier, is a "feature" referring to the full-length sequence (e.g. 150 nucleotides) or a k-mer (e.g. 8-mers)?

  2. How is a taxonomy classified into seven classes (kingdom, phylum, family, ...). Is the multinomial Naive Bayes classifier applied at each node? I am trying to understand how the hierarchal classification was done.

  3. Can I get a dumbed-down explanation of what a "target" is? I am thinking a target is the taxonomy(e.g. k__Bacteria; p__Cyanobacteria; c__Synechococcophycideae; o__Synechococcales; f__Synechococcaceae; g__Synechococcus; s__) but I am not too sure. Maybe answers to the previous two questions will help me out.

Thank you all for all your help so far!