# Exponential random graph models

Exponential random graph models (ERGMs) are a family of statistical models for analyzing data about social and other networks.[1] Examples of networks examined using ERGM include knowledge networks,[2] organizational networks,[3] colleague networks,[4] social media networks, networks of scientific development,[5] and others.

## Background

Many metrics exist to describe the structural features of an observed network such as the density, centrality, or assortativity.[6][7] However, these metrics describe the observed network which is only one instance of a large number of possible alternative networks. This set of alternative networks may have similar or dissimilar structural features. To support statistical inference on the processes influencing the formation of network structure, a statistical model should consider the set of all possible alternative networks weighted on their similarity to an observed network. However because network data is inherently relational, it violates the assumptions of independence and identical distribution of standard statistical models like linear regression.[8][9] Alternative statistical models should reflect the uncertainty associated with a given observation, permit inference about the relative frequency about network substructures of theoretical interest, disambiguating the influence of confounding processes, efficiently representing complex structures, and linking local-level processes to global-level properties.[10] Degree-preserving randomization, for example, is a specific way in which an observed network could be considered in terms of multiple alternative networks.

## Definition

The Exponential family is a broad family of models for covering many types of data, not just networks. An ERGM is a model from this family which describes networks.

Formally a random graph ${\displaystyle Y\in {\mathcal {Y}}}$ consists of a set of ${\displaystyle n}$ nodes and ${\displaystyle m}$ dyads (edges) ${\displaystyle \{Y_{ij}:i=1,\dots ,n;j=1,\dots ,n\}}$ where ${\displaystyle Y_{ij}=1}$ if the nodes ${\displaystyle (i,j)}$ are connected and ${\displaystyle Y_{ij}=0}$ otherwise.

The basic assumption of these models is that the structure in an observed graph ${\displaystyle y}$ can be explained by a given vector of sufficient statistics ${\displaystyle s(y)}$ which are a function of the observed network and, in some cases, nodal attributes. This way, it is possible to describe any kind of dependence between the undyadic variables:

${\displaystyle P(Y=y|\theta )={\frac {\exp(\theta ^{T}s(y))}{c(\theta )}},\quad \forall y\in {\mathcal {Y}}}$

where ${\displaystyle \theta }$ is a vector of model parameters associated with ${\displaystyle s(y)}$ and ${\displaystyle c(\theta )=\sum _{y'\in {\mathcal {Y}}}\exp(\theta ^{T}s(y'))}$ is a normalising constant.

These models represent a probability distribution on each possible network on ${\displaystyle n}$ nodes. However, the size of the set of possible networks for an undirected network (simple graph) of size ${\displaystyle n}$ is ${\displaystyle 2^{n(n-1)/2}}$. Because the number of possible networks in the set vastly outnumbers the number of parameters which can constrain the model, the ideal probability distribution is the one which maximizes the Gibbs entropy.[11]

## References

1. ^ Harris, Jenine K (2014). An introduction to exponential random graph modeling. ISBN 9781452220802. OCLC 870698788.
2. ^ Brennecke, Julia; Rank, Olaf (2017-05-01). "The firm's knowledge network and the transfer of advice among corporate inventors—A multilevel network study". Research Policy. 46 (4): 768–783. doi:10.1016/j.respol.2017.02.002. ISSN 0048-7333.
3. ^ Harris, Jenine K (2013). "Communication Ties Across the National Network of Local Health Departments". AMEPRE American Journal of Preventive Medicine. 44 (3): 247–253. doi:10.1016/j.amepre.2012.10.028. ISSN 0749-3797. OCLC 4937103196. PMID 23415121.
4. ^ Brennecke, Julia (2019). "Dissonant Ties in Intraorganizational Networks: Why Individuals Seek Problem-Solving Assistance from Difficult Colleagues". AMJ Academy of Management Journal. ISSN 0001-4273. OCLC 8163488129.
5. ^ Harris, Jenine K; Luke, Douglas A; Shelton, Sarah C; Zuckerman, Rachael B (2009). "Forty Years of Secondhand Smoke Research. The Gap Between Discovery and Delivery". American Journal of Preventive Medicine. 36 (6): 538–548. doi:10.1016/j.amepre.2009.01.039. ISSN 0749-3797. OCLC 6980180781. PMID 19372026.
6. ^ Wasserman, Stanley; Faust, Katherine (1994). Social Network Analysis: Methods and Applications. ISBN 978-0-521-38707-1.
7. ^ Newman, M.E.J. (2003). "The Structure and Function of Complex Networks". SIAM Review. 45 (2): 167–256. arXiv:cond-mat/0303516. Bibcode:2003SIAMR..45..167N. doi:10.1137/S003614450342480.
8. ^ Contractor, Noshir; Wasserman, Stanley; Faust, Katherine (2006). "Testing Multitheoretical, Multilevel Hypotheses About Organizational Networks: An Analytic Framework and Empirical Example" (PDF). Academy of Management Review. 31 (3): 681–703. doi:10.5465/AMR.2006.21318925.
9. ^ Harris, Jenine K (2014). An introduction to exponential random graph modeling. ISBN 9781452220802. OCLC 870698788.
10. ^ Robins, G.; Pattison, P.; Kalish, Y.; Lusher, D. (2007). "An introduction to exponential random graph models for social networks". Social Networks. 29 (2): 173–191. doi:10.1016/j.socnet.2006.08.002. hdl:1959.3/216571.
11. ^ Newman, M.E.J. (2010-03-25). "Other Network Models". Networks. pp. 565–585. ISBN 978-0-19-920665-0.