Learning Based Approach for Search Engine Selection in Metasearch

In recent years, the web has become a huge source of information, which is mostly unstructured in the form of text or image. However, every search engine uses their own method or algorithm for ranking the retrieved results. The main goal of Metasearch over the single search engine is increased coverage and a consistent interface to ensure that result from several places can be meaningfully combined. In this paper we propose a learning based Query similarity using rank merge list of document approach for search engine selection algorithms to identify the most useful search engines that are likely to contain the relevant documents for the user query. The objective of search engine selection is to improve efficiency as it would result in sending a query to potentially useful underlying search engines only. Finally, it concludes the paper by pointing out some open issues and possible direction of future research related to search engine selection.