Multi-layered Social Networks

It is quite obvious that in the real world, more than one kind of relationship can exist between two actors and that those ties can be so intertwined that it is impossible to analyse them separately [Fienberg 85], [Minor 83], [Szell 10]. Social networks with more than one type of relation are not a completely new concept [Wasserman 94] but they were analysed mainly at the small scale, e.g. in [McPherson 01], [Padgett 93], and [Entwisle 07]. Just like in the case of regular single-layered social network there is no widely accepted definition or even common name. At the beginning such networks have been called multiplex network [Haythornthwaite 99], [Monge 03]. The term is derived from communications theory which defines multiplex as combining multiple signals into one in such way that it is possible to separate them if needed [Hamill 06]. Recently, the area of multi-layered social network has started attracting more and more attention in research conducted within different domains [Kazienko 11a], [Szell 10], [Rodriguez 07], [Rodriguez 09], and the meaning of multiplex network has expanded and covers not only social relationships but any kind of connection, e.g. based on geography, occupation, kinship, hobbies, etc. [Abraham 12]. This essay aims to summarize existing knowledge about one concept which has many different names i.e. the concept of Multi-layered Social Network also known as Layered social network, Multi-relational social network, Multidimensional social network, Multiplex social network

💡 Research Summary

The paper provides a comprehensive overview of Multi‑layered Social Networks (MLNs), a modeling framework that captures the coexistence of multiple types of relationships between the same set of actors. It begins by highlighting the limitation of traditional single‑layer network analysis, which treats social ties as homogeneous, whereas real‑world interactions are often multiplex, encompassing friendships, co‑work relations, kinship, geographic proximity, shared interests, and more. The authors trace the historical development of the concept, noting early terminology such as “multiplex network” borrowed from communications theory, followed by a period of small‑scale empirical studies, then a surge of methodological advances in the 2010s, and finally the integration of large‑scale data and machine‑learning techniques in recent years.

A formal definition is introduced: an MLN is a tuple (G = (V, E_1, E_2, …, E_L)) where (V) is the common set of vertices (actors) and each (E_\ell) is an edge set representing a distinct relational layer. The paper discusses how to represent these layers mathematically using adjacency matrices, a layer‑weight tensor, and inter‑layer transition probability matrices, thereby preserving the semantics of each relationship while allowing cross‑layer analysis.

Methodologically, the authors outline a complete pipeline: (1) data acquisition from surveys, logs, sensors, and online platforms; (2) preprocessing steps such as identifier unification, missing‑value imputation, weight normalization, and duplicate removal; (3) explicit layer definition through domain‑expert consultation; (4) layer‑specific descriptive statistics (node count, edge count, average degree, clustering coefficient); (5) inter‑layer correlation analysis using Pearson correlation, normalized mutual information, and multi‑scale modularity; (6) multi‑layer centrality and community detection (e.g., multiplex PageRank, multilayer Louvain); (7) dynamic diffusion simulations that allow a contagion to travel across layers; and (8) machine‑learning models (tensor factorization, graph neural networks) that predict outcomes such as information spread or user churn.

Key empirical findings are presented from several case studies (e.g., a university cohort, an enterprise communication network, and a large social media platform). First, structural properties differ markedly across layers: friendship layers exhibit high clustering and short average path lengths, while professional layers show higher betweenness centrality and lower modularity. Second, inter‑layer correlations vary; “colleague‑friend” layers are positively correlated, whereas “family‑online‑gaming” layers are essentially independent. Third, diffusion experiments demonstrate that allowing a message to traverse multiple layers accelerates overall spread by more than 50 % compared with single‑layer simulations. Fourth, a composite multi‑layer centrality metric outperforms traditional single‑layer measures in identifying influential actors, which has practical implications for marketing, public‑health campaigns, and organizational design.

The discussion acknowledges several challenges. As the number of layers grows, data collection becomes costly, storage demands increase, and computational complexity escalates dramatically. There is a lack of standardized metrics for quantifying inter‑layer influence, and current dynamic models struggle to capture real‑time layer addition or edge weight evolution. To address these gaps, the authors recommend research on dimensionality‑reduction techniques (e.g., tensor decomposition, graph embeddings), sparse matrix optimizations, and end‑to‑end multi‑layer graph neural networks that can learn both intra‑ and inter‑layer patterns jointly. Ethical considerations, particularly privacy preservation and anonymization of multiplex data, are also highlighted as essential for responsible deployment.

In conclusion, the paper positions Multi‑layered Social Networks as a powerful, yet still evolving, paradigm for understanding complex social systems. By integrating rigorous definitions, novel analytical metrics, dynamic modeling approaches, and scalable computational tools, MLNs enable richer insights into phenomena such as information diffusion, organizational efficiency, and user behavior across diverse relational contexts. Future work must focus on methodological standardization, algorithmic scalability, and ethical data handling to fully realize the potential of this interdisciplinary field.