Anytime Measures for Top-k Algorithms

Top-k queries on large multi-attribute data sets are fundamental operations in information retrieval and ranking applications. In this paper, we initiate research on the anytime behavior of top-k algorithms. In particular, given specific top-k algorithms (TA and TASorted) we are interested in studying their progress toward identification of the correct result at any point during the algorithms’ execution. We adopt a probabilistic approach where we seek to report at any point of operation of the algorithm the confidence that the top-k result has been identified. Such a functionality can be a valuable asset when one is interested in reducing the runtime cost of top-k computations. We present a thorough experimental evaluation to validate our techniques using both synthetic and real data sets.

• Type: Conference
• Year: 2007
• Where: VLDB
• Authors: Benjamin Arai, Gautam Das, Dimitrios Gunopulos, Nick Koudas
• Keywords: Data Sets, Database, Specific Top-k Algorithms, Top-k Algorithms, VLDB 2007