Efficient k-word proximity search

In this paper, we propose the ldquoaddedrdquo use of proximity search to a Web search query for narrowing down the set of documents returned as answers to a keyword based search query. This approach adds value to Web search query results by allowing users to better express what they are looking for. Most of the current search engines provide limited proximity search behaviour such as allowing only two query terms. While there are many algorithms for k-word near proximity search, there is no work for k-word ordered proximity search. This paper presents (a) a new algorithm for k-word ordered proximity search in a document, which runs in O(nlogk) time per document where n is the number of words (terms) in the document, and k is the number of query terms in a query, (b) enhancements to ranking techniques related to proximity search, and (c) a suggestion involving frequent combinations of query terms in order to help users locate the desired documents more accurately.