An Effective Index for Uncertain Data
Chih Wu Chung, Ching Hung Pan, Chuan Ming Liu · 2014
As the modern technologies advance, one can easily access or acquire data and the amount of a data collection increases. The data derived in the emerging computing environments are also usually incomplete or uncertain, such as the sensed data which may not be accurate due to the transmission loss or the error on the sensing devices. How to manage and process the large amount of uncertain data becomes challenging. One of the important approaches for managing data is indexing. In this paper, we propose an effective index structure, US+-tree, for uncertain data in terms of number of I/Os. US+-tree can support point query, range query, top-k query and probability nearest neighbor query. In comparison with the existing MV-tree and US-tree, US+-tree performs the best due to a shorter tree height and less number of internal nodes. We also perform an extensive simulated experiment for validating the proposed index structure.