次へ: (1)Generalization by Meaning Granularity
上へ: Irreducible Minimum Vector
戻る: Irreducible Minimum Vector
In order to reduce the dimension, we use "generalization methods."
Generalization is an inductive inference method to find rules from
examples. In this paper, "generalization" means degenerating the
semantic attributes which have little effect in information retrieval
into the upper attribute. The weight of the attributes which are deleted
from bases are added to the weight of the upper node.
Here, we consider the following two kinds of generalizations:
Jin'ichi Murakami
平成13年10月5日