A Document Summary Method Based on the Relative Spatial Length between Sentences

摘要

This work presents a document summary method which utilizes the relative position between sentences to generate summaries of documents with different levels of details for users. Sentence sets with different spatial distances centered around the searched keywords are extracted and used to generate the summaries. Afterward, we discuss the results of a feasibility study carried out on Japanese news articles, highlighting how this approach could be used to generate summaries with different levels of details.