Ranking of Closeness Centrality for Large-Scale Social Networks
read more
Citations
A Critical Review of Centrality Measures in Social Networks
Influence analysis in social networks: A survey
Deep Recursive Network Embedding with Regular Equivalence
Content-based filtering for recommendation systems using multiattribute networks
Identifying all-around nodes for spreading dynamics in complex networks
References
A note on two problems in connexion with graphs
Centrality in social networks conceptual clarification
The anatomy of a large-scale hypertextual Web search engine
The Anatomy of a Large-Scale Hypertextual Web Search Engine.
Probability Inequalities for sums of Bounded Random Variables
Related Papers (5)
Frequently Asked Questions (12)
Q2. What have the authors stated for future works in "Ranking of closeness centrality for large-scale social networks" ?
There are many directions to extend this study. – Second, the condition under which the heuristic algorithm results in the least number of SSSP computation is an open problem and would be quite interesting to study. – Fourth, the authors would like to study the stability of their algorithms for top-k query using random sampling. This may be possible for some classes of social networks with certain properties on their average distance distributions.
Q3. What are the measures of centrality commonly used in network analysis?
Four measures of centrality that are widely used in network analysis are degree centrality, betweenness centrality4, closeness centrality, and eigenvector centrality5.
Q4. What is the common way to represent a social network?
A social network is typically represented as a graph, with individual persons represented as vertices, the relationships between pairs of individuals as edges, and the strengths of the relationships represented as the weights on edges (for the purpose of finding the shortest weighted distance, the authors can treat lower-weight edges as stronger relationships).
Q5. how many samples are used to calculate the average distance of a vertex?
If the algorithm uses ` = α log n²2 samples (α > 1 is a constant number) which will cause the probability of ²∆ error at each vertex to be bounded above by 1n2 , the probability of ²∆ error anywhere in the graph is then bounded from above by 1n (≥ 1 − (1 − 1n2 )n).
Q6. What is the key in reducing the running time of the algorithm?
The key in reducing the running time of the algorithm is to find the right sample size ` to minimize `+ |E|, the total number of SSSP calculations.
Q7. What is the closeness centrality of v?
The closeness centrality cv of vertex v [Be65] is defined ascv = n− 1Σu∈V d(v, u) . (2.1)In other words, the closeness centrality of v is the inverse of the average (shortestpath) distance from v to any other vertex in the graph.
Q8. What is the way to rank vertices?
The authors first provide a basic ranking algorithm TOPRANK(k), and show that under certain conditions, the algorithm ranks all top k highest closeness centrality vertices (with high probability) in O((k + n 2 3 · log 13 n)(n log n + m)) time, which is better thanAx = λx, where λ is the greatest eigenvalue of A to ensure that all values xi are positive by the Perron-Frobenius theorem.
Q9. What is the closestness centrality of a graph?
The problem to solve in this paper is to find the top k vertices with the highest closeness centralities and rank them, where k is a parameter of the algorithm.
Q10. How does the algorithm calculate the closeness centrality?
In [EW04], Eppstein and Wang developed an approximation algorithm to calculate the closeness centrality in time O( log n²2 (n log n + m)) within an additive error of ²∆ for the inverse of the closeness centrality (with probability at least 1 − 1n ), where ² > 0 and ∆ is the diameter of the graph.
Q11. How many vertices are correlated with the closeness measure?
Their study shows that, with a 50% sample of the original network nodes, average correlations for the closeness measure ranged from 0.54 to 0.71.
Q12. What is the closestness centrality of a vertex?
Since the average shortest-path distances are bounded above by ∆, the average difference in average distance (the inverse of closeness centrality) between the ith-ranked vertex and the (i + 1)th-ranked vertex (for any i = 1, . . . , n− 1) is O(∆n ).