AGENTS '98 Proceedings of the second international conference on Autonomous agents
Foundations of statistical natural language processing
Foundations of statistical natural language processing
The "DGX" distribution for mining massive, skewed data
Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining
Statistical mechanics of complex networks
Statistical mechanics of complex networks
RCV1: A New Benchmark Collection for Text Categorization Research
The Journal of Machine Learning Research
The slashdot zoo: mining a social network with negative edges
Proceedings of the 18th international conference on World wide web
On the evolution of user interaction in Facebook
Proceedings of the 2nd ACM workshop on Online social networks
Controversial users demand local trust metrics: an experimental study on Epinions.com community
AAAI'05 Proceedings of the 20th national conference on Artificial intelligence - Volume 1
Power-Law Distributions in Empirical Data
SIAM Review
KONECT: the Koblenz network collection
Proceedings of the 22nd international conference on World Wide Web companion
Hi-index | 0.00 |
This paper presents several measures of fairness and inequality based on the degree distribution in networks, as alternatives to the well-established power-law exponent. Networks such as social networks, communication networks and the World Wide Web itself are often characterized by their unequal distribution of edges: Few nodes are attached to many edges, while many nodes are attached to only few edges. The inequality of such network structures is typically measured using the power-law exponent, stating that the number of nodes with a given degree is proportional to that degree taken to a certain exponent. However, this approach has several weaknesses, such as its narrow applicability and expensive computational complexity. Beyond the fact that power laws are by far not a universal phenomenon on the Web, the power-law exponent has the surprising property of being negatively correlated with the usual notion of inequality, making it unintuitive as a fairness measure. As alternatives, we propose several measures based on the Lorenz curve, which is used in economics but rarely in networks study, and on the information-theoretical concept of entropy. We show in experiments on a large collection of online networks that these measures do not suffer under the drawbacks of the power-law exponent.