Rule-based Approach for Arabic Root Extraction: New Rules to Directly Extract Roots of Arabic Words

Extracting word roots in Arabic language is very prob-lematic due to the specific morphological and structural changes in the language. To address this problem, several techniques have been proposed. This paper continues the problem of identifying and exploiting relationship amongst Arabic letters for Arabic root extraction begun in [1]. Eight different rules that detect the root letters according to other letters in the word have been pro-posed and tested, four of them benefiting from the idea of morphological substitution (MUTATION). The ap-proach has been evaluated using the Holy Quran words. The evaluation results show a promising root extraction algorithm