Automatic Abstracting Based On Semantic Web

With the progress of society and development of technology, electronic text information emerges in plenty. In order to obtain keynote of text, automatic abstraction arises spontaneously with the advantage of easiness and rapidness. The thesis designs and realizes the automatic abstracting system manipulating English text based on semantic web.The thesis is based on the traditional statistic-based automatic abstraction adding the semantic analysis which makes typical ontology-WordNet in the semantic web as core. It not only makes abstraction free to field constraints but also contains semantic analysis, it has seven steps through the process to manipulate awaiting text, the seven steps are text unit processing, word frequency count, anaphora resolution, vector space model building, sentence importance count, sentence similarity count and post processing.I have made picking-up automatic abstraction experiment in the abstraction system with lots of articles, comparing with picking up abstraction using "automatic summarization" in office word. It has been proved that the automatic abstraction using this automatic abstracting system is high accuracy and easy to understand according to the internal and external assessment.