JIDR: Towards building hybrid Arabic stemmer

The Arabic Language Processing knew these last decades a massive rise, giving birth to specialized tools like: information retrieval systems, text categorization, orthographic correctors, generators of words, automatic summarizers, etc... However the development of such tools strongly relies on of a certain number of basic modules such as Stemming, which consists of converting each word to into canonical form. The aim of this paper is to present our system JIDR a stemmer of Arabic text. Where the originality of the work, resides in the fact of cohabiting three different techniques (dictionary, removing affix and morphological analysis). But also, in the architecture set up to treat weak verbs.