New root-based stemming approach for Arabic Language

Abstract EN

The Arabic language is expanding in the world day after day.

The presence of the Arabic language on the internet grew around 6.091% in the last fifteen years, it is the highest growth of the ten top online languages, and the number of Arabic documents increases rapidly.

Therefor it necessity to improve the one of important tools (stemming) for the Arabic Information Retrieval (IR) techniques.

Many researchers agree on the benefits of stemming to enhance the efficiency in information retrieval system.

In this paper we presents a new (root-based) stemming approach for Arabic language this technique is based on prefix and suffix removal and matching with Arabic patterns.

The implementation and evaluation of this stemmer shows the improvement in the accuracy relative to other algorithms.