In order to create a good mode, we need to annotate with consistency and need to have enough training data for each entity and relation. Currently there is no tool or view to see what kind of words and how many words were annotated for each entities/relations. If we could have such view, we can see which entity/relation has less training data and so we can add more training data or vice-versa. Also with such view, we can check the consistency of annotations, making sure that same words are not annotated to same entity.
Why is it useful?
|Who would benefit from this IDEA?||Human Annotators|
How should it work?