How the annotation of this DocumentAnnotator should be printed as extra information after a one-word-per-line (OWPL) format.
How the annotation of this DocumentAnnotator should be printed as extra information after a one-word-per-line (OWPL) format. If there is no document annotation, return the empty string. Used in Document.owplString.
Given a sequence of MentionPairLabels for a document, compute features of the pair and return both
Given a sequence of MentionPairLabels for a document, compute features of the pair and return both
Given the index of a mention, create positive and negative labels for this mention and its prodecessors
Given the index of a mention, create positive and negative labels for this mention and its prodecessors
Mentions for this document
Index of current mention to generate labels for
Training Labels for this Mention
Generate the labels used for training
Generate the labels used for training
This is expected to be the true coreference class for the document
Sequence of training labels for this document
Find each mentions best scoring antecedent.
Find each mentions best scoring antecedent. If the antecedent has a cluster add the new mention if not, create a new entity and add both mentions Currently does not create singleton entities
Expects nontarget coref class that is pre annotated with mentions
Store head words which are seen over a default 20 times in the model
Store head words which are seen over a default 20 times in the model
Documents to generate counts from
Types of Pairs Pruned during Training
Types of Pairs Pruned during Training
How the annotation of this DocumentAnnotator should be printed in one-word-per-line (OWPL) format.
How the annotation of this DocumentAnnotator should be printed in one-word-per-line (OWPL) format. If there is no per-token annotation, return null. Used in Document.owplString.