Characterizing how 'distributional' NLP corpora distance metrics are

Open in new window