Attribute-Based and Value-Based Clustering: An Evaluation.

摘要: In most research on concept acquisition from corpora, concepts are modeled as vectors of relations extracted syntactic structures. the case modifiers, these often specify values attributes, in (attr red); this is unlike what typically proposed theories knowledge representation, where defined terms their attributes (e.g., color). We compared models based with using lexical clustering basis for comparison. find that attribute-based work better than value-based ones, and result shorter descriptions; but mixed including both best all.