AR2Concept Automatic extraction concepts from Arabic text language

Abstract

Our objective is the design and realization of an automatic system of extraction of concepts from a text in the Arabic language as a first step towards the creation of ontology. The architecture we adopt is an original approach for Arabic language texts that combines the semantic concept extraction method based on the Latent Semantic Analysis documentary search technique with the K-means algorithm. Faced with the problem posed by the K-means algorithm for the number of clusters to be fixed, we propose a solution that we have evaluated on a set of texts. The first results are satisfactory. Our AR2Concept system allowed the identification of concepts with an f-measure rate of around 80%