Finding Skyline Nodes in Large Networks

Arijit Khan, Vishwakarma Singh, Jian Wu

Abstract: Given a query node q in some graph dataset, we say that a node u is not dominated by another node v, if (1) the distance between u and q is less than the distance between v and q, or (2) the set of query node labels carried by u is a superset of the set of query node labels carried by v. A node is called a skyline node with respect to some query node, if it is not dominated by any other node in the graph, except the query node. Given a query node, the set of skyline nodes are more interesting, because any non-skyline node is farther from the query node and also it carries less number of query node labels, compared to some skyline node. Since there can be many such skyline nodes, they are usually ranked based on how many nodes each of them dominates. In this paper, we propose efficient techniques to determine the top-k skyline nodes from large network datasets.