Document decomposition for contextual indexation

  • Mohamed Salim EL BAZZI et al.


The systems of features extraction from texts use a wide range of different approaches and
techniques. On the one hand, this is due to the wide morphosyntactic range that the textual
document possesses and the various problems that may arise during the extraction of knowledge.
On the other hand, Text Mining and Understanding is very promising, and approaches and
methods that can be used as precise tools still gaining momentum. In this paper, we are interested
in sentence classification techniques to generate contexts. The investment in sentence grouping
approaches provides great precision for indexing large documents. We validate our approach with
metrics on the results of document classification