Title :
Arabic text summarization using Rhetorical Structure Theory
Author :
Ibrahim, Ahmed ; Elghazaly, Tarek
Author_Institution :
Dept. of Comput. & Inf. Sci., Cairo Univ., Cairo, Egypt
Abstract :
The Rhetorical Structure Theory (RST) is a descriptive theory of a major aspect of the structure of natural text. It is applied in English as well as other languages such as, French and Japanese but there are still no clear efforts to apply RST in Arabic. This paper provides a framework to apply RST in Arabic, in order to improve the ability of extracting the semantic behind the text. First, by hypothesizing rhetorical relations and gathering quantitative and qualitative analyses for all relations that are correctly defined through this framework. Secondly, using these relations to identify the text parts are very important in order to extract informative summaries from the whole text. Finally, framework results scored 26% recall, 34% precision and 29% F-measure.
Keywords :
natural language processing; text analysis; Arabic text summarization; informative summary extraction; rhetorical structure theory; Educational institutions; Informatics; Joints; Natural language processing; Presses; Satellites; RST; Rhetorical Structure Theory;
Conference_Titel :
Informatics and Systems (INFOS), 2012 8th International Conference on
Conference_Location :
Cairo
Print_ISBN :
978-1-4673-0828-1