Efficient processing of top [formula omitted] group skyline queries

Abstract

For a given multi-dimensional data set, a group skyline query returns the optimal groups not dominated by any other group of equal size. The group skyline query is a powerful tool in many applications that call for optimal groups. However, it is common to return a large number of results which make users overwhelmed since it prevents them from making quick and rational decisions. To address this problem, we first identify and formulate a top k group skyline (TkGSky) query which returns k optimal groups dominating the highest number of points in the given data set. Next, new pruning strategies are presented to reduce the search space. Then, we propose efficient algorithms by exploiting novel techniques including a grouping strategy, a hybrid strategy, and a point-based replacement strategy, respectively. Finally, we also develop an approximate algorithm to further improve the TkGSky query performance. The performance of the proposed algorithms is studied by extensive experiments over synthetic and real datasets.