Indexing method for transitive relationships of product information

Article type: Research Article

Authors: Leea, Hyunja^{; ;} | Shimb, Junho^;

Affiliations: Laboratoire Le2i, Faculté des Sciences, Université de Bourgogne, 21078, Dijon Cedex, France | Department of Computer Science, Sookmyung Women's University, Yongsan-gu, Seoul, 140-742, Korea

Note: [] This paper is an extended version of the authors' previous work [9].

Note: [] This work was performed when she was a doctoral student at Sookmyung Women's University. She is currently a postdoc at Université de Bourgogne, France.

Note: [] Corresponding author. E-mail: [email protected].

Abstract: To successfully use a relational database management system (RDBMS) as a repository for product information, the RDBMS must efficiently process and properly answer ontological queries. The key to processing the ontological queries is whether the various semantic relationships among the concepts of the product ontology are likewise well-processed. In particular, the transitive relationships (e.g., is-a, component-of relationships) such as ancestors-descendents, parents-children, and taxonomy of products must be processed successfully. We propose an efficient index using a numbering scheme (labeling scheme) to process queries over transitive relationships.

Keywords: Product information, e-commerce, index, relational database

DOI: 10.3233/WIA-2011-0223

Journal: Web Intelligence and Agent Systems: An International Journal, vol. 9, no. 4, pp. 351-362, 2011