Characterizing World Wide Web search strategies

Abstract

<p>The World Wide Web (or just "Web") is a hypermedia system operating on the Internet which provides access to Web pages that exist on thousands of servers throughout the world. The growth of the Web has been phenomenal, with approximately 100 million pages currently available. To find relevant information in this huge distributed database without being overwhelmed by information overload requires strategies which make appropriate use of the logical links among these Web pages, along with the directories and search engines which have been developed to support Web search. We describe five search objectives: exploration, known­ item search, topic search, general search, and exhaustive search. Web search strategies are then described for finding the most relevant information in each situation and thus meet these objectives in the most effective manner. A model is described which can be used to estimate the total number of relevant Web pages on a particular topic, given the search results and overlaps in retrievals from two or more search engines on that topic. This model is particularly useful in exhaustive search mode, and an example ofits use is demonstrated with a real Web search that uses the results from five search engines.</p>