multinet-package: Multilayer social network analysis and mining
In multinet: Analysis and mining of multilayer social networks

multinet-package

R Documentation

Multilayer social network analysis and mining

Description

This package defines a class to store multilayer networks and functions to pre-process, analyze and mine them.

With multilayer social network we indicate a network where vertices (V) are organized into multiple layers (L) and each node corresponds to an actor (A), where the same actor can be mapped to nodes in different layers. Formally, a multilayer social network as implemented in this package is a graph G = (V, E) where V is a subset of A x L.

In this manual, multinet.IO describes functions to read and write multilayer networks from/to file and the file format. To quickly test some features of the library, some existing multilayer networks are also included (multinet.predefined). A synthetic multilayer network can be generated using the growing models described in multinet.generation.

Updating and getting information about the basic components of a multilayer network (layers, actors, vertices and edges) can be done using the methods described in multinet.properties, multinet.update and multinet.edge_directionality. multinet.navigation shows how to retrieve the neighbors of a node. Attribute values can also be attached to the basic objects in a multilayer network (actors, layers, vertices and edges). Attribute management is described in multinet.attributes.

Each individual layer as well as combination of layers obtained using the data pre-processing (flattening) functions described in multinet.transformation can be analyzed as a single-layer network using the iGraph package, by converting them as shown in multinet.conversion. We can also visualize small networks using the method described in multinet.plotting and the layouts in multinet.layout.

Multilayer network analysis measures are described in multinet.actor_measures (for single-actor, degree-based measures), multinet.distance (for measures based on geodesic distances) and multinet.layer_comparison (to compare different layers).

Communities can be extracted using various clustering algorithms, described in multinet.communities.

Most of the methods provided by this package are described in the book "Multilayer Social Networks". These methods have been proposed by many different authors: extensive references are available in the book, and in the documentation of each function we indicate the main reference we have followed for the implementation. For a few methods developed after the book was published we give specific references to the corresponding literature.