Web Retrieval with Query Expansion: A Parallel Retrieving Technique

Abstract

Most people consider that the World Wide Web (WWW) is a mine of information. The explosive growth in the WWW, not only in the amount of information but also in contents of Web pages, makes traditional search engines inadequate approach to the retrieval of documents or web pages that are most relevant to user needs (degree of relevance) in a short time. To improve the information retrieval process, from both time and degree of relevance to user need, parallel genetic algorithms could be utilized. In this paper, island genetic algorithm (IGA) is utilized to achieve parallelism and speed up the web information ‎retrieval process. Four different islands with different selection methods and fitness functions are suggested to be used to improve degree of relevance. To achieve parallel behavior, the four islands are executed independently on different ‎servers. Query expansion technique is used to add useful words to user query and enhance number of retrieved documents. ‎Finally, the results obtained by the four islands are combined and passed to a decision making phase to select the documents most pertinent ‎to user needs. Cosine similarity measure is used to evaluate the performance of the proposed ‎technique.‎