Probabilistic Reverse Top-k Query on Probabilistic Data

Reverse top-k queries have received much attention from research communities. The result of reverse top-k queries is a set of objects, which had the k-most interest based on their objects' references. Moreover, answering the queries on probabilistic data has been studied in many applications. The most common problem with uncertain queries is how to calculate their probabilities. Currently, there are some proposed solutions for selecting answers to queries and calculating probabilistic values based on users' preferences. In this paper, we study answering reverse top-k queries on probabilistic data. Firstly, we propose a novel method to calculate probabilistic tuples based on the expected theory. Secondly, we present the advantages of our approach against the traditional approach. Furthermore, we upgrade the new algorithm using two techniques of R-tree and upper-lower bound. The experimental results illustrate the efficiency of the proposed algorithm compared to the traditional algorithms in terms of scalability.