Domain-specific keyphrase extraction and near-duplicate article detection based on ontology

The significant increase in number of the online newspapers has given web users a giant information source. The users are really difficult to manage content as well as check the correctness of articles. In this paper, we introduce algorithms of extracting keyphrase and matching signatures for near-duplicate articles detection. Based on ontology, keyphrases of articles are extracted automatically and similarity of two articles is calculated by using extracted keyphrases. Algorithms are applied on Vietnamese online newspapers for Labor & Employment. Experimental results show that our proposed methods are effective.