Study on Domain-Dependent Automatic Text Summarization

Studies ways of generating text summary by utilizing domain knowledge to dissolve the limitation that statistical summary is defective in understanding document meanings.Discusses methods in constructing the corpus and knowledge base;proposes methods to postprocess the summary by generalizing summary sentences and adding knowledge to the summary.Developes text summarizing software based on the theory presented.In comparing it with MS Word 2003,it is seen that the approaches are effective and efficient;and can be extended to other domains.