A Song – Author Relationship Extractor for Vietnamese news articles.

Abstract

Relationship extraction is a vital task which is required for many research fields such as Ontology construction or Q - A systems. Good results have been reached for widely used languages such as English while the achievement for Vietnamese is still limited. This paper presents a relationship extractor for Vietnamese which is trained with semi-supervised learning. A set of metrics is employed to qualify extraction result. And a simple NER system is integrated to boost the extractor's efficiency