Probabilistic MaxRS Queries on Uncertain Data

Abstract

Given a set of spatial objects with scores and a size of a rectangle, MaxRS (Maximizing Range Sum) queries retrieve the location of the rectangle which maximizes the sum of the scores of all objects covered by the rectangle. MaxRS queries can be employed in many useful applications such as finding an attractive area for tourists. So far, some literatures proposed efficient algorithms for MaxRS query processing in traditional databases. In real environments, however, values of data are inherently uncertain. Therefore, for the first time, we address the problem of processing probabilistic MaxRS (P-MaxRS) queries. P-MaxRS queries retrieve a set of tuples $\langle l, p\rangle$, where p is the probability that l is a MaxRS location. Our algorithm prunes locations with zero probability to be MaxRS, and then efficiently calculates p for each location where can be a MaxRS location. Our experiments demonstrate the efficiency of our algorithm.