Abstract
Statistical tables have usually been classified by using many classes for meeting users' potential needs. The number of classes is often excessive for an individual user who has a need to extract particular data. The statistical tables are cross tabulated by two or more classes, but the format for the tables is not unified. Furthermore, there are tables that contain classes with hierarchical structures. For these reasons, it is not easy for a user to extract statistical values of classes from already published statistical tables. This paper presents a new notation method of the statistical tables. This method employs Wang's notation method and describes the paths from attributes of non-uniform format containing hierarchical classes to statistical values by XML. In order to confirm the effectiveness of this notation method, we carried out experiments using model tables and actual medical statistical tables which the Ministry of Health, Labor and Welfare of Japan had published. It is shown that the proposed notation worked well for easy extraction of particular data. A new system that changes the statistical tables into the proposed notation automatically, and that enables users to extract particular data from the changed statistical data interactively, is also developed.
