An efficient multi-agent system combining POS-Taggers for arabic texts

摘要: In this paper, we address the problem of Part-Of-Speech tagging Arabic texts with vowel marks. After description specificities language and induced difficulties on task POS-tagging, propose an approach combining several methods. One these methods, based sentences patterns, is original very attractive. We present, afterward, multi-agent architecture that adopted for conception realization our POS-tagging system. The justified by need collaboration, parallelism competition between different agents. Finally, expose implementation evaluation system implemented.