Finding Compositional Skyline Based on Sharing Strategy
Leigang Dong, Guohua Liu · 2017
Traditional skyline queries are mainly to return the best points from a large data set. However, it only identifies the individual points. In this paper, we perfectly present the concept of compositional skyline(C-Skyline), which returns the best results from all the compositions with k points. It is more useful in big data mining and cloud computing. In order to compute C-Skyline efficiently, we present two sharing strategies which show how to find candidate compositions, and we build a dominance graph to reflect dominance relations among the points in adjacent levels. Then we propose two heuristic algorithms to query C-Skyline compositions: the ordinary algorithm and the improved algorithm. Based on the synthetic dataset and the real NBA dataset, the experiments manifest the abundant information of C-Skyline, and the efficiency of our algorithms.