Closeness Centrality on Uncertain Graphs

Centrality is a family of metrics for characterizing the importance vertex in graph. Although large number centrality have been proposed, majority them ignores uncertainty graph data. In this paper, we formulate closeness on uncertain graphs and define batch evaluation problem that computes subset vertices an We develop three algorithms, MS-BCC , MG-BCC MGMS-BCC based sampling to approximate specified vertices. All these algorithms require perform breadth-first searches (BFS) starting from sampled possible worlds To improve efficiency exploit operation-level parallelism BFS traversals simultaneously execute shared sequences operations searches. Parallelization realized at different levels algorithms. The experimental results show proposed can efficiently accurately given faster than both because it avoids more repeated executions operation traversals.