An Efficient Algorithm for Probabilistic RkNN Query on Uncertain Data with Large k

Abstract

Recently, the query on uncertain data attracts much attention and it is great significance for probabilistic reverse k neighbor query on uncertain data based on location-based services (LBS). However, the relevant research is less and immature. Probabilistic reverse k nearest neighbor (PRkNN) requests the query point of reverse k neighbor query and the probability is greater than the given threshold. The main problem of the existing research is that, when the value k is larger, a reduction of the query efficiency is obvious. In this paper, we propose an algorithm called PRCLU for PRkNN with larger k, including pruning phase and verification phase. The pruning phase with a minimum circle to enclose the uncertain data, which performs pruning with the region, then followed by probabilistic pruning strategy in sequence. The results of the experiment show that the algorithm PRCLU is better than other similar methods when k is larger.