An application of pattern matching stemmer in Arabic dialogue system

Abstract

This paper proposes the building of a stemmer for the Arabic language. This stemmer is largely based on pattern matching and pattern strength techniques. Stemmers are algorithms to extract root from a word by removing its affixes. Stemming has been applied for large number of applications, such as: indexing, information retrieval systems, and web search engines. This paper will also proposes the application of stemming as a pre-processing stage in a dialogue system (DS). The proposed stemmer was compared with three other well known stemmers and achieved favourable accuracy. © 2011 Springer-Verlag.