Arabic Words Root Extraction Using Modified Successor Variety

Abstract

The process of returning words to their linguistic roots is one of the most important stages of preparation for computerizing the Arabic text, this process aims to: Overcoming the multiple morphological forms of the Arabic language, rooting algorithms are used in many applications of language processing, such as information retrieval applications, machine translation, text summarization, optical character recognition, and other applications. In this research, a new algorithm was proposed to extract roots of the Arabic words, depending on the adopted successor variety algorithm, by changing the body of the proposed text in this algorithm, so that it consists of several morphological formulas associated with the word to be rooted, the proposed algorithm has been tested on approximately six thousand derived words from 137 deferent roots, with 44 morphological forms for each root, the ratio of correct extracted roots using the proposed algorithm was about 98.5%.