Snowboot: Bootstrap Methods for Network Inference

Abstract

Complex networks are used to describe a broad range of disparate social systems and natural phenomena, from power grids to customer segmentation to human brain connectome. Challenges of parametric model specification and validation inspire a search for more data-driven and flexible nonparametric approaches for inference of complex networks. In this paper we discuss methodology and R implementation of two bootstrap procedures on random networks, that is, patchwork bootstrap of Thompson et al. (2016) and Gel et al. (2017) and vertex bootstrap of Snijders and Borgatti (1999). To our knowledge, the new R package snowboot is the first implementation of the vertex and patchwork bootstrap inference on networks in R. Our new package is accompanied with a detailed user's manual, and is compatible with the popular R package on network studies igraph. We evaluate the patchwork bootstrap and vertex bootstrap with extensive simulation studies and illustrate their utility in application to analysis of real world networks. The accepted manuscript is published in the R Journal. This is an open access article under the CC BY 4.0 license (https://creativecommons.org/licenses/by/4.0/). (The R Journal (2018) 10:2, pages 95-113; doi: https://doi.The new R package snowboot (Ramirez-Ramirez et al., 2018) is the first to offer vertex and patchwork bootstrap in R. Moreover, to our knowledge, there currently exist only two more R packages implementing bootstrap analysis of networks, bootnet and sna. The package bootnet assesses uncertainty in estimation of edge-weights and centrality indices but does not account for network dependence structure among vertices. The package sna implements parametric bootstrapping of edges to generate new random graphs that can be further used for hypothesis testing of matching between the observed network and randomized baseline network as well as canonical correlation coefficients. This paper aims to further fill the gap and showcase utility of nonparametric resampling for network inference, with a particular focus on vertex and patchwork bootstraps. snowboot imports The R Journal Vol. XX/YY, AAAA 20ZZ

Crawling subsampling for multivariate spatial autoregression model in large-scale networks

“…Thus, to estimate the parameters in these models, extending the CS method could be an interesting research problem. Secondly, for the parameters in different network models and the properties of the network, the bootstrap estimator of the parameters deserves a separate study; see [35,15,8] for further discussions. Thirdly, we have regarded the network structure and nodal features as fixed during the observation period.…”

Pitfalls of Climate Network Construction—A Statistical Perspective

“…While most studies directly perform bootstrapping on the graph structure and not on underlying data (Chen et al 2018;Levin and Levina 2019), a similar idea has been previously suggested (Friedman et al 1999). To the best of our knowledge, it has not yet been applied in climate science.…”

Block dense weighted networks with augmented degree correction

“…There have been a few attempts to incorporate the bootstrap into the network context because the bootstrap is an important, general purpose statistical tool, but one whose i.i.d. assumptions are violated in most network contexts (Chen et al, 2018;Levin & Levina, 2019;Bhattacharyya & Bickel, 2015;Lunde & Sarkar, 2019;Chen & Onnela, 2019;Green & Shalizi, 2022;Li et al, 2020;Krivitsky et al, 2011;Desmarais & Cranmer, 2012b). Our proposal is closer to the parametric bootstrap of (Schmid & Desmarais 2017), which generates bootstrapped networks using MPLE ERGM parameters of a single observed network to estimate valid confidence intervals of ERGM parameters.…”