Portable Text Summarization

Background

Summarization approaches are often divided into two main groups, text extraction and text abstraction. Text abstraction is in many aspects similar to what humans abstractors do when writing an abstract, even though professional abstractors often utilize surface-level information such as headings, key phrases and position in the text as well as the overall organization of the text into more or less genre specific sections (Liddy 1991, Endres-Niggemeyer et al. 1995, Cremmins 1996). The parsing and interpretation of text is a venerable research area that has been investigated for many years. In this area we have a wide spectrum of techniques and methods ranging from word by word parsing to rhetorical discourse parsing as well as more statistical methods, or a mixture of all. Also the generation of text is a vigorous research field with techniques ranging from canned text and template filling to more advanced systems with discourse planners and surface realizers.