Automated Keyword Extraction using Support Vector Machine from Arabic News Documents

Abstract

Keyword extraction is an indispensable step for many natural language processing and information retrieval applications such as; text summarization and search engine optimization. Keywords hold the most important information describing the content of a document. With the increasing volume and variety of unlabeled documents on the Internet, the need for automatic keyword extraction methods increases. Even though keyword extraction can be used in many applications, Arabic research in the field still lacking. In this paper, a supervised learning technique that uses statistical features and Support Vector Machine classifier was implemented and applied to extract the keywords from Arabic news documents. The proposed supervised learning approach achieved a precision of 0.77 and a recall of 0.58.