The problem of author identification can be defined in the following way: given a set of documents written by an author and a new document, we have to decide if this last one was written or not by the same author as the other documents. For solving this problem we have suggested and implemented various approaches: counting method, vote technique and supervised learning which explore several models of document representation. The experiences carried out using the collections of PAN-CLEF 2014 challenge have confirmed the interest of our approaches as well as their performance in terms of running time.