A review of dynamic network models with latent variables

Area of Science:

• Network Science
• Statistical Modeling
• Machine Learning

Background:

• Dynamic networks are complex systems where relationships evolve over time.
• Understanding the underlying structure of these networks is crucial for various applications.
• Traditional static models often fail to capture the temporal dynamics inherent in network data.

Purpose of the Study:

• To provide a selective review of statistical modeling techniques for dynamic networks.
• To focus on latent variable models, including latent space and latent class models (stochastic blockmodels).
• To discuss applications and identify open problems in the field.

Main Methods:

• Overview of static network models.
• Introduction and discussion of dynamic extensions of latent variable models.
• Review of existing literature on applications and data sources.

Main Results:

• Comprehensive summary of latent space and latent class models for dynamic networks.
• Identification of diverse applications across different domains.
• Compilation of key challenges and open research questions.

Conclusions:

• Latent variable models offer powerful tools for analyzing dynamic network structures.
• Further research is needed to address current challenges in dynamic network modeling.
• The review provides a foundation for future advancements in understanding evolving network phenomena.