Extensibility in Arabic full-text indexing. (c2003)

Abstract

The Internet has become a gateway for an abundant resource for information. Web search engines provide a useful mechanism through which we can look and access information. Most of the research has been dedicated to Latin languages. Using the Latin alphabet, this work proposes an extensible mechanism for indexing and Arabic search. The method expands the capabilities of Arabic search engines and indexing engines. In this work we propose a rules compiler that extends the Arabic indexing rules used by the indexer at runtime. We validate our approach using a prototype that was built using Java and MS SQL Server as a backend RDBMS. Tests have been made on a sample of Arabic documents and a starting set of indexing rules.