All the Web, Alta Vista and Google : a study of the retrieval performance of three web search engines.

The purpose of this essay is to perform an evaluation regarding retrieval performance of three web search engines. The search engines included are All the Web, Alta Vista and Google. These have been tested using ten queries within different subject areas chosen by the author. The first twenty hits for each question have been analyzed for relevancy. Relevancy is judged on a three-point scale; 0, 0,5 and 1 where 0 goes to irrelevant and inactive or duplicate documents, 0,5 point goes to partially relevant documents and 1 point goes to highly relevant documents. Criteria for these relevance judgments are formulated as to prevent hesitation and partiality. The measure used is precision. When calculating precision special methods are adopted to give credit to those search engines that present relevant hits early in the ranked lists. The number of duplicate, inactive and mirror links are also recorded for each search engine. The results show that Google is the best performing search engine of these three concerning precision. All the Web is the second best while Alta Vista performs worst of the three. All the Web has the largest number of duplicate links while Google has the least. None of the search engines have significant problems with inactive or mirror links. It is the author' s hope that more specified and unified methods for evaluating web search engines will arise since it would increase the possibility of comparing different results.Uppsatsnivå: