Rule merging in a rule-based Arabic stemmer

Abstract

Semitic languages require more complicated systems for processing their morphology. Arabic language, for example, exhibits a very complex but very regular morphological structure. Many approaches were proposed to analyze Arabic language at the morphological level. Proposed approaches can be classified into table lookup, linguistic, combinatorial and rule-based techniques.This paper proposes a new approach to enhance a rule-based Arabic stemmer. The enhancement is based on rule merging process to reduce number of rules, increase language coverage and maintain the same level of performance.