backbone: An R package to extract network backbones

Abstract

Networks are useful for representing phenomena in a broad range of domains. Although their ability to represent complexity can be a virtue, it is sometimes useful to focus on a simplified network that contains only the most important edges: the backbone. This paper introduces and demonstrates a substantially expanded version of the backbone package for R, which now provides methods for extracting backbones from weighted networks, weighted bipartite projections, and unweighted networks. For each type of network, fully replicable code is presented first for small toy examples, then for complete empirical examples using transportation, political, and social networks. The paper also demonstrates the implications of several issues of statistical inference that arise in backbone extraction. It concludes by briefly reviewing existing applications of backbone extraction using the backbone package, and future directions for research on network backbone extraction.

Assessing the impact of structural modifications in the construction of surveillance network for transboundary animal diseases: the role of backbone and sentinel nodes

“…Several methods have been used to extract the backbone from the network. In this study, we used the LD (Local Degree) approach applied by Neal et al (11), which assesses edge importance by considering the endpoints of the nodes and prioritizing those linked to central nodes. Following edge ranking and normalization, the most significant edges of each node are retained in a sparser network that, nevertheless, preserves hierarchical structures.…”

“…Identifying likely transmission paths and extracting subgraphs of the network where most information spreads are essential for studying epidemic-spreading phenomena and designing intervention strategies (9). Indeed, the network backbone is the core component obtained by filtering all the redundant information while maintaining the network hierarchy (8)(9)(10)(11). Nodes of the backbone are potential candidates for surveillance systems.…”

signnet: An R package for analyzing signed networks

“…The resulting networks will thus be weighted. Several methods exist to turn a weighted projection into an unweighted network where only the most significant edges are included (Neal, 2022). Projecting signed networks is not as straightforward, because "nullification" of edges can occur.…”

On network backbone extraction for modeling online collective behavior

“…Some works propose specific rules to the selection of salient edges, e.g., as in [ 80 , 81 ] where salient edges are identified based on the idea of overlapping communities. A large number of previous studies has proposed and analyzed methods for extracting backbones in bipartite networks [ 37 , 38 , 82 ]. Considering a bipartite structure consisting of artifacts and agents, a family of methods, such as the Fixed Sequence Degree Method (FSDM), the Fixed Fill Method (FFM), the Fixed Row Method (FRM), the Fixed Column Method (FCM), and the Stochastic Degree Sequence Model (SDSM), identify salient edges by constraining the degree sequence of either artifacts (FCM), agents (FRM), both (FSDM and SDSM), or neither (FFM).…”

“…A second challenge in the study of collective behavior is the lack of ground truth data, as it is often the case for studies on online social media. Although information about the phenomenon of interest may be present in some scenarios [37,38], the lack of ground truth prevents the evaluation of the quality of the extracted backbone, making the comparison and choice of methods hard [39][40][41][42]. Authors thus resort to the evaluation of topological metrics, such as community modularity, density, clustering coefficient of the extracted backbone, which more clearly defines sub-structures than the original network [20,37,[43][44][45].…”