Difference between revisions of "Text summarization"
From Cohen Courses
Jump to navigationJump to searchPastStudents (talk | contribs) |
PastStudents (talk | contribs) |
||
| Line 12: | Line 12: | ||
== Evaluation == | == Evaluation == | ||
| − | One commonly used evaluation metric in summarization is [[ | + | One commonly used evaluation metric in summarization is [[ROUGE]], which is used in NIST's Document Understanding Conferences' summarization tasks. It is considered as an [[Automatic Evaluation Method]]. |
== Example Systems == | == Example Systems == | ||
Latest revision as of 15:45, 30 November 2010
Contents
Summary
Text Summarization (also known as summarization, and automatic summarization) is a natural language processing problem which focuses on creating shortened versions of texts with computer algorithms/software that retain the important points of the original piece of text.
Common Approaches
Common approaches to text summarization can typically be classified into one of the following categories:
- Extraction, extracts most important information (sentences or paragraphs) from original text and copies them to make summary
- Abstraction, paraphrases sections in the original text and relies on language generation to make the summaries coherent
Evaluation
One commonly used evaluation metric in summarization is ROUGE, which is used in NIST's Document Understanding Conferences' summarization tasks. It is considered as an Automatic Evaluation Method.
Example Systems
References / Links
- A bit outdated website with some references related to text summarization - [1]
- Wikipedia article on automatic summarization - [2]
