Online estimating the k central nodes of a network (original) (raw)

This paper addresses the problem of estimating the k most central nodes in a network with limited online sampling. By exploring the impact of sampling and identification errors, it develops algorithms that utilize random walk sampling to effectively identify influential nodes across various real-world networks. The findings suggest that random walks can significantly capture a high fraction of central nodes early in the sampling process, while highlighting the challenges posed by different degree distributions and correlations among centrality measures.