An Effective Document Search Technique by Semantic Relationship Approach

Abstract

Office applications are becoming a major pillar of today's organizations since they are used to edit a vast amount of digital documents. Finding these office documents in large databases that fit users' needs is becoming increasingly important. Traditional search tools that employ keyword and phrase matching between the query and search index alone tend to offer high recall and low precision. The search users are faced with too many irrelevant results. In order to solve this problem, we propose a novel search system that effectively searches the target documents by the search query whose definition is based on the document type, search terms and the semantic relationship between the search terms and the target documents. We present a technique that collects search terms and their semantic relationship from the documents of some office applications to generate the XML-based search indices that can effectively locate the office documents. Furthermore, our query optimization algorithm is also presented.