An incremental hypercube approach for finding best matches for vague queries

Abstract

In this article we present an extension of the Vague Query System (VQS) which allows the user to efficiently find the best matching record for ad-hoc queries. The VQS operates on top of existing database systems. It maps arbitrary types of attributes to the Euclidean space in order to represent semantic background information. Due to the multi tier concept of the VQS we can not apply conventional multidimensional search methods directly and so we have chosen to use an iterative approach. The concept works on the basis of an incremental extension of the search intervals around the query values which is repeated until the best match is proven to be found. As an indexing method for effectively accessing the semantic background information a slightly modified version of the pyramid technique (Berchtold et. al.) is applied.