Skip to main content
. 2018 May 18;46(Web Server issue):W523–W529. doi: 10.1093/nar/gky428

Figure 4.

Figure 4.

Interactive learning workflow. New documents are first manually annotated or refined after applying the pre-trained or string match tagger. Interactive learning follows three steps interactively: (i) Training TaggerOne using annotated documents, (ii) tagging biological concepts in new documents using the trained model and (iii) refining the documents by correcting annotation mistakes.