TOWARDS AUTO-CONSTRUCTION OF DOMAIN ONTOLOGY: AN AUTO-CONSTRUCTED DOMAIN CONCEPTUAL LEXICON AND ITS APPLICATION TO EXTRACTIVE SUMMARIZATION

摘要

A domain conceptual lexicon can be used for the construction of a domain ontology.In this paper, a method of auto-constructing a domain conceptual lexicon is proposed.The conceptual lexicon is represented by WordNet senses and their relations.Using the domain conceptual lexicon, an extractive summarization method is proposed.A characteristic of the scheme is that each sentence of a document is represented by a set of WordNet senses and constitutes a fuzzy transaction for mining conceptual lexicon and ranking relevance.A prototype of this automatic text summarization scheme is built and an intrinsic method with the information-retrieval criteria is used for measuring the summary quality.