Hierarchical categorization is a powerful and convenient method so that it
is commonly used in various areas, for example ontologies and information
categorization. Although each hierarchy is useful, there are problems to
manage multiple hierarchies. In this paper, we propose an alignment method
between concept hierarchies by using a statistical method. By using this
method, a concept that exists in one hierarchy system but does not in the
other can be located in a suitable position in the other. The key idea is to
find similar categories between two systems to be able to transfer concepts
from one system to the other. Similarity is measured by "k(kappa) statistic"
based on instances belonging categories. The experiments of our method with
concept hierarchies of Yahoo! and LYCOS result over 80% of accuracy to
estimate appropriate positions of concepts between two hierarchies.