Odaies: ontology-driven adaptive Web information extraction system

This paper proposes an ontology-driven self-adapting approach in the semi-structured Web information extraction field, where ontology provides semantic support and plays a central role during the extraction process. It outperforms traditional wrapper systems in adaptiveness and maintenance. Firstly, we build a domain-dependant ontology. Then we design three templates generating algorithms, which have self-adaptiveness and self-maintenance based on the ontology, to perform Web page information extraction. Experiment results show that our prototype system can achieve 100% recall and 97.64% precision.