Abstract:Full probability models are critical for the statistical modeling of complex networks, and yet there are few general, flexible and widely applicable generative methods. We propose a new family of probability models motivated by the idea of network growth, which we call the Latent Order Logistic (LOLOG) model. LOLOG is a fully general framework capable of describing any probability distribution over graph configurations, though not all distributions are easily expressible or estimable as a LOLOG. We develop inferential procedures based on Monte Carlo Method of Moments, Generalized Method of Moments and variational inference. To show the flexibility of the model framework, we show how so-called scale-free networks can be modeled as LOLOGs via preferential attachment. The advantages of LOLOG in terms of avoidance of degeneracy, ease of sampling, and model flexibility are illustrated. Connections with the popular Exponential-family Random Graph model (ERGM) are also explored, and we find that they are identical in the case of dyadic independence. Finally, we apply the model to a social network of collaboration within a corporate law firm, a friendship network among adolescent students, and the friendship relations in an online social network.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to propose a new probabilistic model—the Latent Order Logistic (LOLOG) model—to address the lack of a general, flexible, and widely applicable generative method in the statistical modeling of complex networks. Specifically, the paper focuses on the following points: 1. **Limitations of Existing Models**: - One of the most popular generative models currently is the Exponential-family Random Graph Model (ERGM). Although ERGM is flexible and general, it suffers from poor scalability and model degeneracy issues. - Other specific network generation processes, while capable of describing certain specific features, often lack an inference framework for parameter estimation from observed data and sometimes do not even have a fully specified probabilistic model. 2. **Advantages of the LOLOG Model**: - **Avoiding Degeneracy**: The LOLOG model has advantages in avoiding model degeneracy. - **Ease of Sampling**: The LOLOG model is easy to generate samples from. - **Model Flexibility**: The LOLOG model framework is very flexible and can describe the probability distribution of any graph configuration. 3. **Application Examples**: - The paper demonstrates the flexibility and practicality of the LOLOG model by applying it to the collaboration network of corporate law firms, the friendship network among adolescent students, and the friendship relationships in online social networks. ### Summary By proposing the LOLOG model, the paper fills the gap in the statistical modeling of complex networks, which lacks a general, flexible, and widely applicable generative method. The LOLOG model not only has theoretical advantages but also demonstrates its effectiveness and flexibility in practical applications.

A New Generative Statistical Model for Graphs: The Latent Order Logistic (LOLOG) Model

New network models facilitate analysis of biological networks

A Latent Logistic Regression Model with Graph Data

Exponential-family Random Network Models

Maximum Likelihood Latent Space Embedding of Logistic Random Dot Product Graphs

Emergent Probability - A directed Scale-Free Network Approach to Lonergan's Generic Model of Development

Latent Space Generative Model for Bipartite Networks

GraphOpt: Learning Optimization Models of Graph Formation

Null models for network data

Fast and scalable likelihood maximization for Exponential Random Graph Models with local constraints

Community Detection through Likelihood Optimization: In Search of a Sound Model

Generalized Latent Factor Models for Social Network Analysis

Detecting Latent Communities in Network Formation Models

A Probabilistic Generative Model for Latent Business Networks Mining.

On the geometry of exponential random graphs and applications

New Specifications for Exponential Random Graph Models

An introduction to exponential random graph (p*) models for social networks

On Graphical Models via Univariate Exponential Family Distributions

Graph-Sparse Logistic Regression

Latent Space Approaches to Social Network Analysis

A Latent Moving Average Model for Network Regression