A Survey on Automatic Summarization

Abstract

With the increasing popularity of Internet and the diversity of information obtaining technologies, the amount of quickly growing information has gone beyond our imaginations. Many techniques were presented to help users to find the desired information from large data set quickly and accurately, automatic summarization is an effective approach. In this paper, after careful investigation, existing automatic summarization techniques are classified as five categories: automatic extraction, understanding-based automatic summarization, information extraction, automatic summarization based on discourse and automatic summarization based on user-query. The history of automatic summarization is outlined. The principles of the five categories methods are respectively described in detail. In the end, the five categories methods are compared and the future work is discussed.