WARNING: REPOSITORY UNDER CONSTRUCTION !! First complete release expected during May 2024, stay tuned ! Or follow @GZamora_Lopez in Twitter.
… is a generalised perspective for the study of graphs and complex networks from the viewpoint of dynamical systems, instead of the traditional combinatorial approach. Many existing graph and network metrics are based on different assumptions of how information, particles, agents or signals may propagate along a network. However, for many of these, the underlying dynamical model and its assumptions are hidden, or have been ignored. SRNA is a paradigm shift in the way we study complex networks and interpret their results: from a data-driven (combinatorial) tradition to a (dynamical) model-based data analysis perspective. The goal is to take full advantage of model-based network analyses with transparency at the front. To encompass existing metrics into an integrated methodology and to provide analysis tools in which the underlying assumptions are recognized from the beginning, thus favouring the interpretability of results.
This repository is the entry point for interested users of Stimulus-Response Network Analysis (SRNA) and its corresponding Python library SiReNetA . Here we provide a set of tutorials and examples––in the form of Jupyter notebooks––to help interested users walk through the basics and specific examples.
- 1_Whatis_Sireneta : A short overview of the main ideas and goals of stimulus-response network analysis.
- 2_Canonical_Models : Description of simple (canonical) models for various classes of dynamical propagation on networks, following a variety of assumptions and constraints.
- …
- Getting Started and Overview : Installing SiReNetA and a first example using Stimulus-Response Network Analysis.
- Response to Stimulus and Calculating Metrics : Fundamentals of Stimulus-Response Network Analysis. Undertanding the stimulus-responses at all levels (pair-wise, node-wise and network level). Illustration and interpretation of pair-wise responses for undirected and directed chains. Global-responses, time-to-peak / time-to-saturation distances. NOT THERE YET, BUT NEEDED??? Comparison of TTP for leaky-cascade and the continuous diffusion (as used in Arnaudon's paper).
- Canonical Propagation Models : Presentation of the different canonical models.
- Comparing Networks : Use-Case. Illustration of how renormalization of connectivity allows for network comparison under SRNA, without the need of generating surrogates.
- Network Distance : In-Depth. Generalization of the (geodesic) graph distance between nodes as response times. Validity for weighted networks.
- Weighted Networks : Use case to show the natural ability of SRNA to deal with weighted networks.
- Graph metrics from dynamics : In-Depth. Demonstration of how classical (combinatorial) graph metrics are derived from a dynamical perspective.
See resources in folder Reproduce_Papers/.
The documentation of the SiReNetA library can be accessed 'online' as usual for other python packages. Typing either help(module_name) or module_name? in an IPython interactive session will display the docstring of the module. Equally, module_name.func_name? will show the usage instructions for each function.
- G. Zamora-López and M. Gilson An integrative dynamical perspective for graph theory and the analysis of complex networks. Chaos 34, 041501 (2024).
- M. Gilson, N. Kouvaris, et al. Network analysis of whole-brain fMRI dynamics: A new framework based on dynamic communicability. NeuroImage 201, 116007 (2019).
- M. Gilson, N. Kouvaris, G. Deco & G.Zamora-Lopez "Framework based on communicability and flow to analyze complex networks" Phys. Rev. E 97, 052301 (2018).
List of recommendations and further topics that should be covered in the tutorials.
- How to properly interpret the concepts and the metrics.