next up previous
次へ: (2) Generalization by Weight 上へ: Irreducible Minimum Vector 戻る: Minimum set of Vector

(1) Generalization by Granularity

The granularity of the meanings represented by a literal word is smaller than that of a semantic attribute. Accordingly, the granularity of a Semantic-Vector is rough compared to a Word-Vector. And in accordance with the generalization by granularity of semantic attributes, the performance of information retrieval will decrease. Then, while generalizing the bases, the change in performance is traced to find the minimum set of bases.



Jin'ichi Murakami 平成13年10月5日