École Nationale Supérieure de
Techniques Avancées

Résumé

More than 80 % of the world’s information is stored in a natural language text form which is highly unstructured and merely understandable for machines. During the last decade, Researchers have invested in Information Extraction field in order to allow computers to understand resources contained in web pages, so they can interact actively with users. Several approaches focus on binary relations by extracting data from text and presenting it in a triple format. Yet, extracting complex information from causal , contrasting, opinion sentences presents a big challenge for most of these systems. In this paper, we survey existing approaches to date by presenting an overview of their strengths and weaknesses and finally we suggest promising directions of future work.