Efficient Processing of Probabilistic Threshold Top-k Queries Based on X-tuple in Uncertain Database

Top-k queries are widely used in analyzing and processing uncertain data. In the uncertain database, the xtuple consists a number of alternatives which are mutually exclusive, the independence still remains among the x-tuples. Probabilistic threshold top-k queries (PT-k queries) return the tuple taking a probability of at least threshold p to be in the top-k list, but it didn´t take alternatives in x-tuple as a whole so that there are limits on its usable range. We define a novel queries, probabilistic threshold x-tuple top-k queries (PT-x-k queries), which solves that problem. Meanwhile, the paper provides pruning methods for the algorithm. The great efficiency of our algorithm is proved by experiments on various data sets.