Indexation based on the textual evolution

In this work, we have studied different methods that try to bring out the notion of textual evolution. The primary objective of these methods is to bring additional information figuring the logical process adopted during the text creation. The first method presented in this paper relies on an automaton for creating images. It also describes the rules of the machine and chosen evaluation measures. The second method is based on statistical techniques. The fusion of these two methods results in a nonsupervised classification method to achieve better results on different corpuses.