Weighted Multiplex Network Analysis
Weighted Multiplex Network Analysis (Multi-Layer Network Analysis with Edge Weights) · Also known as: WMNA, weighted multilayer network analysis, weighted multi-relational network analysis, multiplex weighted graph analysis
Weighted multiplex network analysis studies systems in which the same set of actors are connected through multiple types of relationships simultaneously, and each relationship carries a quantitative strength or frequency. By capturing both the variety and the intensity of ties across layers, it reveals patterns invisible to single-layer or unweighted network approaches.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use weighted multiplex network analysis when your data contain the same actors connected by multiple distinct relationship types AND those ties have measurable strengths or frequencies. Typical candidates include social networks with several interaction channels (e.g., email, in-person, collaboration), biological networks with gene-regulation and protein-interaction layers carrying expression levels, or trade networks across multiple commodity types with monetary volumes. Do not use it when you have only one type of relationship (use standard weighted network analysis instead), when edge weights are absent or unreliable (use unweighted multiplex analysis), or when the node sets across layers are substantially different (use multilayer heterogeneous network methods).
Strengths & limitations
- Captures both relational diversity (multiple tie types) and tie intensity simultaneously, yielding a richer structural portrait than single-layer or unweighted approaches.
- Layer-specific centrality and community measures allow comparison of how a node's role differs across relationship contexts.
- Aggregate metrics (overlapping strength, participation coefficient) identify nodes that are universally influential versus those dominant in only one layer.
- Flexible framework: layers can represent time periods, relationship categories, or spatial scales, making the method broadly applicable.
- Compatible with standard weighted-network algorithms (Dijkstra paths, Louvain community detection) extended to the multiplex setting.
- Data requirements are steep: complete, reliable weight information is needed for every layer, which is often unavailable in practice.
- The number of parameters and diagnostics grows with the number of layers, complicating interpretation and reporting.
- No single agreed-upon software implementation covers all weighted multiplex metrics; analyses may need to combine multiple packages.
- Layer boundary decisions (what counts as a separate layer) are researcher-driven and can substantially affect results.
Frequently asked
How is weighted multiplex analysis different from ordinary weighted network analysis?
Ordinary weighted network analysis uses a single edge-weight matrix for one type of relationship. Weighted multiplex analysis maintains a separate weight matrix for each relationship type (layer) and adds cross-layer metrics — such as participation coefficient and overlapping strength — that capture how a node's weighted connectivity profile varies across layers. The extra information reveals whether a node's importance is layer-specific or universally robust.
Do the weights in different layers need to be on the same scale?
Not necessarily, but comparisons across layers require normalisation. A common approach is to divide each layer's weight matrix by its maximum or mean weight so that layers contribute comparably to aggregate metrics. Skipping normalisation can make high-weight layers dominate aggregate scores regardless of structural importance.
What software can I use?
Python libraries muxViz and py-multinet, as well as R packages multinet and igraph (with manual layer loops), cover most weighted multiplex metrics. NetworkX can handle individual layers; custom code is often needed to combine them into supra-adjacency matrices.
How many layers is too many?
There is no fixed upper limit, but with more than about ten layers the combinatorial diagnostics become unwieldy and interpretability suffers. When many layers are highly correlated, dimensionality reduction of layers (e.g., tensor decomposition) is advisable before applying full multiplex metrics.
Can I use weighted multiplex analysis with temporal data?
Yes. One common approach treats each time window as a separate layer, making the multiplex temporal. This overlaps with temporal network analysis; the choice depends on whether time-window identity or relationship-type identity is the primary analytical lens.
Sources
- Battiston, F., Nicosia, V., & Latora, V. (2014). Structural measures for multiplex networks. Physical Review E, 89(3), 032804. DOI: 10.1103/PhysRevE.89.032804 ↗
- Kivela, M., Arenas, A., Barthelemy, M., Gleeson, J. P., Moreno, Y., & Porter, M. A. (2014). Multilayer networks. Journal of Complex Networks, 2(3), 203-271. DOI: 10.1093/comnet/cnu016 ↗
How to cite this page
ScholarGate. (2026, June 3). Weighted Multiplex Network Analysis (Multi-Layer Network Analysis with Edge Weights). ScholarGate. https://scholargate.app/en/network-analysis/weighted-multiplex-network-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Multiplex Network AnalysisNetwork analysis↔ compare
- Weighted Betweenness CentralityNetwork analysis↔ compare
- Weighted Community DetectionNetwork analysis↔ compare
- Weighted Eigenvector CentralityNetwork analysis↔ compare
- Weighted Network Diffusion AnalysisNetwork analysis↔ compare