構造化文書とオブジェクトの統合データベース管理システム

概要

論文の詳細を見る
By managing documents in databases, it is possible to associate character strings in documents with other data in databases. The association in most of such efforts is based on pattern matching with strings. Our research focuses on semantics denoted by strings in documents, and provides link mechanism based on this semantics to develop deep integration between document data and other data in databases. The semantics denoted by strings is that i). data being managed by databases and; ii). data possible to be managed by databases. We will call, in this paper, such data as legacy data. By utilizing the link mechanism, there will be the following advantages. ・In addition to traditional pattern matching facilities provided by Information Retrieve Systems (IRS), it becomes possible to retrieve documents based on relationships among legacy data denoted by character strings. ・Furthermore, it becomes possible to utilize database facilities such as functions and indices of some types. Considering the foregoing link mechanism, we will model documents as "text of which some substrings have associated legacy data" instead of "a linear sequence of character strings". Our approach is the introduction of a new ADT named paratext into extensible database. Paratext has a simple data structure with functions having efficient expressive power. The paratext has following features: ・The links are created over instances in databases. ・The links are created by authors of documents when documents are created. ・Legacy data and paratext data are integrated, that is, both are managed in the "same" database. The integrated database proposed in this paper can answer , for example, the following query: Q1: Retrieve articles in which something related to Canada is mentioned, and they ranking scores. In section 2., we will describe the paratext ADT and other ADTs incorporated into our database. Section 3. illustrate the query language in our database. We will also sketch a novel sort dizzy path.
社団法人情報処理学会の論文
1997-09-24

構造化文書とオブジェクトの統合データベース管理システム

スポンサーリンク

概要

著者

関連論文

スポンサーリンク