New Approaches of Ranking Queries in Uncertain Databases

Abstract

New applications such as sensor data monitoring and mobile device tracking, rise up the issue of uncertain data management. Compared to ―certain‖ data, the data in the uncertain database are not exact points, which, instead, often reside within a region. Here the study ranked queries over the uncertain data. The existing system such as decision making, recommendation raising, and data mining tasks proposes only for certain data. We define several fundamental properties including performance, radius range, unique-range, value-invariance and stability. Robust novel solutions speed up the probabilistic ranked query (PRank) with monotonic preference functions over the uncertain database. Specifically, Introduced two effective pruning methods: Spatial and Probabilistic pruning, these reduce the PRank search space. A special case of PRank with linear preference function is also studied and seamlessly integrate these pruning heuristics into the PRank query procedure. A preference function specified by users, a ranked query retrieves k data objects in the database such that their scores (calculated by the given preference function) are the highest. The proposed system tackles the PRank query processing over the join of two distinct uncertain databases. Extensive experiments will be conducted to demonstrate the efficiency and effectiveness in answering PRank queries.