A System for Evaluation of Arabic Root Extraction Methods

In this article, we present a new application that evaluated the performance of a number of the Arabic root extraction methods. The implemented methods in this system are selected according to a previous classification, where these methods are classified into five categories. We have selected a method for each category. These methods are: Light Stemmer, Arabic Stemming without a root dictionary, MT-based Arabic Stemmer, N-gram based on similarity coefficient and N-gram based on dissimilarity coefficient. This evaluation was conducted on the same terms in a corpus of two thousand words and their roots. These words are taken from Arabic dictionary "Lesan Al-Arab". This application has allowed us to have a first original comparison between the evaluated methods. This system works in two ways: normal and automatic.