Abstract
Network analysis has emerged as a key technique in communication studies, economics, geography, history and sociology, among others. A fundamental issue is how to identify key nodes, for which purpose a number of centrality measures have been developed. This paper proposes a new parametric family of centrality measures called generalized degree. It is based on the idea that a relationship to a more interconnected node contributes to centrality in a greater extent than a connection to a less central one. Generalized degree improves on degree by redistributing its sum over the network with the consideration of the global structure. Application of the measure is supported by a set of basic properties. A sufficient condition is given for generalized degree to be rank monotonic, excluding counter-intuitive changes in the centrality ranking after certain modifications of the network. The measure has a graph interpretation and can be calculated iteratively. Generalized degree is recommended to apply besides degree since it preserves most favourable attributes of degree, but better reflects the role of the nodes in the network and has an increased ability to distinguish among their importance.
Similar content being viewed by others
Notes
Degrees in the component given by the set of nodes \(N^1\) are the same in (N, A) and \((N,A^{\prime })\) due to the condition \(a_{ij} = a^{\prime }_{ij}\) for all \(i,j \in N^1\).
There exists a path from node j to node h due to connectedness. Along this path one should find such \(k,m \in N\), for example, \(k = j\) and \(m = h\) if \(a_{jh} = 1\).
\(n \ge 3\) is necessary to get a meaningful start network.
There exists an appropriately small \(\varepsilon \) satisfying the condition of Theorem 1 for any \(\mathsf {d}\) and \(\mathfrak {d}\).
Nevertheless, the lower rank of node 3 may be explained. Landherr et al. (2010) mention cannibalization and saturation effects, which sometimes arise when an actor can devote less time to maintaining existing relationships as a result of adding new contacts. Similarly, the edges can represent not only opportunities, but liabilities, too. For example, a service provider may have legal constraints to serve unprofitable customers. Investigation of these models is leaved for future research.
The proof is available from the author upon request. See also http://www.math.unm.edu/~loring/links/graph_s09/degreeSeq.pdf.
Superscript (k) indicates the centrality vector obtained after the kth iteration step.
In the limit it corresponds to the average degree of neighbours in \(G^{\prime }\) since \(\lim _{\varepsilon \rightarrow \infty } \beta = 1 / {\mathfrak {d}}\).
References
Avrachenkov KE, Mazalov VV, Tsynguev BT (2015) Beta current flow centrality for weighted networks. In: Thai MT, Nguyen NP, Shen H (eds) Computational social networks, of lecture notes in computer science, vol 9197. Springer, New York
Bavelas A (1948) A mathematical model for group structures. Hum Organ 7(3):16–30
Boldi P, Vigna S (2014) Axioms for centrality. Internet Math 10(3–4):222–262
Bonacich P (1987) Power and centrality: a family of measures. Am J Sociol 92(1):1170–1182
Borgatti SP, Everett MG (2006) A graph-theoretic perspective on centrality. Soc Netw 28(4):466–484
Brin S, Page L (1998) The anatomy of a large-scale hypertextual web search engine. Comput Netw ISDN Syst 30(1):107–117
Chebotarev P (1989) Generalization of the row sum method for incomplete paired comparisons. Autom Remote Control 50(8):1103–1113
Chebotarev P (1994) Aggregation of preferences by the generalized row sum method. Math Soc Sci 27(3):293–320
Chebotarev P (2012) The walk distances in graphs. Discrete Appl Math 160(10):1484–1500
Chebotarev P, Shamis E (1997) The matrix-forest theorem and measuring relations in small social groups. Autom Remote Control 58(9):1505–1514
Chien S, Dwork C, Kumar R, Simon DR, Sivakumar D (2004) Link evolution: analysis and algorithms. Internet Math 1(3):277–304
Csató L (2015) A graph interpretation of the least squares ranking method. Soc Choice Welf 44(1):51–69
Dequiedt V, Zenou Y (2015) Local and consistent centrality measures in networks. https://www.gate.cnrs.fr/IMG/pdf/Dequiedt2015.pdf
Freeman LC (1977) A set of measures of centrality based on betweenness. Sociometry 40(1):35–41
Freeman LC (1979) Centrality in social networks: conceptual clarification. Soc Netw 1(3):215–239
Garg M (2009) Axiomatic foundations of centrality in networks. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1372441
Geršgorin S (1931) Über die Abgrenzung der Eigenwerte einer Matrix. Bulletin de l’Académie des Sciences de l’URSS. Classe des sciences mathématiques et naturelles 6:749–754
González-Díaz J, Hendrickx R, Lohmann E (2014) Paired comparisons analysis: an axiomatic approach to ranking methods. Soc Choice Welf 42(1):139–169
Jackson MO (2010) Social and economic networks. Princeton University Press, Princeton
Katz L (1953) A new status index derived from sociometric analysis. Psychometrika 18(1):39–43
Kitti M (2012) Axioms for centrality scoring with principal eigenvectors. http://www.ace-economics.fi/kuvat/dp79.pdf
Klein DJ (2010) Centrality measure in graphs. J Math Chem 47(4):1209–1223
Landherr A, Friedl B, Heidemann J (2010) A critical review of centrality measures in social networks. Bus Inf Syst Eng 2(6):371–385
Leavitt HJ (1951) Some effects of certain communication patterns on group performance. J Abnorm Soc Psychol 46(1):38
Lindelauf RHA, Hamers HJM, Husslage BGM (2013) Cooperative game theoretic centrality analysis of terrorist networks: the cases of Jemaah Islamiyah and Al Qaeda. Eur J Oper Res 229(1):230–238
Masuda N, Kawamura Y, Kori H (2009) Analysis of relative influence of nodes in directed networks. Phys Rev E 80(4):046114
Masuda N, Kori H (2010) Dynamics-based centrality for directed networks. Phys Rev E 82(5):056107
Meyer CD (2000) Matrix analysis and applied linear algebra. Society for Industrial and Applied Mathematics, Philadelphia
Mohar B (1991) The Laplacian spectrum of graphs. In: Alavi Y, Chartrand G, Oellermann OR, Schwenk AJ (eds) Graph theory, combinatorics, and applications, vol 2. Wiley, New York, pp 871–898
Monsuur H, Storcken T (2004) Centers in connected undirected graphs: an axiomatic approach. Oper Res 52(1):54–64
Neumann C (1877) Untersuchungen über das logarithmische und Newton’sche Potential. B. G. Teubner, Leipzig
Nieminen J (1974) On the centrality in a graph. Scand J Psychol 15(1):332–336
Ranjan G, Zhang Z-L (2013) Geometry of complex networks and topological centrality. Phys A Stat Mech Appl 392(17):3833–3845
Rubinstein A (1980) Ranking the participants in a tournament. SIAM J Appl Math 38(1):108–111
Sabidussi G (1966) The centrality index of a graph. Psychometrika 31(4):581–603
Seeley JR (1949) The net of reciprocal influence: a problem in treating sociometric data. Can J Psychol 3(4):234–240
Wasserman S, Faust K (1994) Social network analysis: methods and applications. Cambridge University Press, Cambridge
Acknowledgments
We are grateful to Herbert Hamers for drawing our attention to centrality measures, and to Dezső Bednay, Pavel Chebotarev and Tamás Sebestyén for useful advices. The research was supported by OTKA grant K 111797 and MTA-SYLFF (The Ryoichi Sasakawa Young Leaders Fellowship Fund) Grant ‘Mathematical analysis of centrality measures’, awarded to the author in 2015.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Proof of Theorem 1
Let \(x_\ell = x_\ell (\varepsilon )(N,A)\) and \(x_\ell ^{\prime } = x_\ell (\varepsilon )(N,A^{\prime })\) for all \(\ell \in N\). It can be assumed without loss of generality that \(x_i^{\prime }-x_i \ge x_j^{\prime }-x_j\). Let \(s = x_i^{\prime }-x_i\), \(t = x_j^{\prime }-x_j\), \(u = \max \lbrace x_\ell ^{\prime }-x_\ell : \ell \in N {\setminus } \{ i,j \} \rbrace = x_k^{\prime } - x_k\) and \(v = \min \{ x_\ell ^{\prime }-x_\ell : \ell \in N {\setminus } \{ i,j \} = x_m^{\prime } - x_m\). It will be verified that \(t \ge u\).
Assume to the contrary that \(t < u\). The difference of equations from (1) for node k gives:
Since \(u = x_k^{\prime } - x_k \ge x_\ell ^{\prime } - x_\ell \) for all \(\ell \in N {\setminus } \{ i,j \}\) and \(u \ge t\), we get
Now take the difference of equations from (1) for node i:
Since \(u \ge x_\ell ^{\prime } - x_\ell \) for all \(\ell \in N {\setminus } \{ i,j \}\), we get
An upper bound for u is known from (2), thus
Take the difference of equations from (1) for node m:
Since \(v = x_m^{\prime } - x_m \le x_\ell ^{\prime } - x_\ell \) for all \(\ell \in N {\setminus } \{ i,j \}\) and \(s \ge t\), we get
The difference of equations from (1) for node j results in:
Since \(v \le x_\ell ^{\prime } - x_\ell \) for all \(\ell \in N {\setminus } \{ i,j \}\), we get
A lower bound for v is known from (5), thus
According to our assumption \(t < u\), therefore from (4) and (6):
Obviously, it does not hold if \(\varepsilon \rightarrow 0\). Now an upper bound is determined for the parameter \(\varepsilon \). After some calculations we get:
Let introduce the notations \(\alpha = \varepsilon \left( a_{m i} + a_{m j} + d_i \right) + \varepsilon ^2 \left[ d_i \left( a_{m i} + a_{m j} \right) - a_{k i} d_j \right] \) and \(y = 1 + 2 \varepsilon a_{ki} + \varepsilon ^2 \left[ 2 a_{k i} \left( a_{m i} + a_{m j} + d_j \right) \right] > 1\). Then the fraction on the left-hand side can be written as \((1 + \alpha ) / (y + \alpha )\). However, \((1 + \alpha ) / (y + \alpha ) > 1 / y\) because \(y + \alpha > 0\) and \(y > 1\), hence
Here \(a_{k i},a_{m i},a_{m j} \le 1\) and \(d_j \le \mathfrak {d}\). As \(x_j^{\prime } - x_i^{\prime } \le x_j - x_i \le \mathfrak {d} - \mathsf {d}\) from boundedness (Proposition 1):
which contradicts to the condition of Theorem 1. \(\square \)
Rights and permissions
About this article
Cite this article
Csató, L. Measuring centrality by a generalization of degree. Cent Eur J Oper Res 25, 771–790 (2017). https://doi.org/10.1007/s10100-016-0439-6
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10100-016-0439-6