Estimating the size of Arabic indexed web content

Abstract

Various initiatives designed to increase Arabic Web content have been undertaken in recent years, and now search engines are reporting that the Arabic portion of Web content has grown relative to the overall Web content. An accurate estimate of Arabic Web content is crucial for those interested in studying it and enriching it. In this paper, we propose a statistics-based system to estimate the size of Arabic indexed Web content using three popular search engines; Google, Yahoo and Bing. Our system relies on selecting sample words from an Arabic corpus to estimate the size of the Arabic Web content indexed by the search engines and the overlap among them. We have used Arabic Wikipedia as a corpus, as it provides diversified content accessed by a large number of Internet users. Our results show that, as of December 2010, the size of the Arabic indexed Web content was estimated at 2 to 2.1 billion pages.