Ego-betweenness centrality in link streams

Abstract

The ability of a node to relay information in a network is often measured using betweenness centrality. In order to take into account the fact that the role of the nodes vary through time, several adaptations of this concept have been proposed to timeevolving networks. However, these definitions are demanding in terms of computational cost, as they call for the computation of time-ordered paths. We propose a definition of centrality in link streams which is node-centric, in the sense that we only take into account the direct neighbors of a node to compute its centrality. This restriction allows to carry out the computation in a shorter time compared to a case where any couple of nodes in the network should be considered. Tests on empirical data show that this measure is relatively highly correlated to the number of times a node would relay information in a flooding process. We suggest that this is a good indication that this measurement can be of use in practical contexts where a node has a limited knowledge of its environment, such as routing protocols in delay tolerant networks.

Temporal walk based centrality metric for graph streams

“…Although many studies tried to identify the best estimates for the importance of a social media user, to the best of our knowledge, there are only two previous studies ( Rozenshtein and Gionis 2016 ; Ghanem et al 2017 ) that propose data stream updateable centrality measures. The algorithm of ( Rozenshtein and Gionis 2016 ), which we analyze in Section Temporal PageRank , cannot incorporate the actual edge arrival times in its calculations.…”

“…We believe our method is superior in using the exact time of interaction between two social media users, resulting in better performance in our prediction task. The algorithm of ( Ghanem et al 2017 ) can be best described as a heuristic version of betweenness centrality to “ego-graphs”, which have paths of length two only. They applied their algorithms for small graphs of less than 250 nodes only.…”

Mining social and cryptocurrency networks

“…In our rst main result, we extend the denition of the Katz index [72] to the edge stream graph computational model where the edges of the network arrive continuously in time. Although many studies tried to identify the best estimates for the importance of a social media user, to the best of our knowledge, only two previous studies [52,131] propose data stream updateable centrality measures.…”

“…Our centrality metric is online updateable, thus ideal for data-intensive applications. To the best of our knowledge, only two previous studies [52,131] propose data stream updateable centrality measures.…”

“…To the best of our knowledge, the only previous such algorithms are temporal PageRank [131] and degree [76] other measures are inecient to update online. For example, in [52], a heuristic version of betweenness centrality was proposed for ego-graphs, which have paths of length two only. They applied their algorithms for small graphs of less than 250 nodes only.…”