Handling ER-topk query on uncertain streams

Abstract

It is critical to manage uncertain data streams nowadays because data uncertainty widely exists in many applications, such as Web and sensor networks. The goal of this paper is to handle top-k query on uncertain data streams. Since the volume of a data stream is unbounded whereas the memory resource is limited, it is challenging to devise one-pass solutions that is both time- and space efficient. We have devised two structures to handle this issue, namely domGraph and probTree. The domGraph stores all candidate tuples, and the probTree is helpful to compute the expected rank of a tuple. The analysis in theory and extensive experimental results show the effectiveness and efficiency of the proposed solution. © 2011 Springer-Verlag.