Bichromatic Reverse Ranking Query in Two Dimensions

Capturing potential customers for a given product based on individual preferences is very important in many personalized applications. Reverse ranking queries are widely employed in this scenario from the perspective of product in database community. Currently, most existing approaches to handle reverse ranking queries generally focus on the d-dimensional space. However, those approaches are oblivious to special properties in the 2-dimensional space which is useful for further optimization. Moreover, there exist many applications, such as data visualization, in the 2-D space. In this work, we propose two general approaches, namely sorting-based method and tree-based pruning method, in order to efficiently process reverse ranking query in the 2-D space. Both methods are able to handle two variants of reverse ranking query i.e., reverse top-k query and top-k reverse query. Analysis and experimental reports on real and synthetic data sets illustrate the efficiency of our proposed methods.