jonnob / rsetse

This package is used to calculate the Strain Elevation Tension Spring embedding (SETSe) for networks in R

rSETSe

Importnant note

The github version of R setse has been updated so that setse_auto_hd has been re-written in c++ using Rcpp Armadillo. This makes it much faster (between 36 -350 times) than the version on CRAN. If you have larger networks I would definately advise using installing from github not CRAN.

An R package for embedding graphs using the SETSe algorithm

This is the R package for the Strain Elevation Tension Spring embeddings (SETSe) algorithm. SETSe is a deterministic graph embeddings algorithm. It converts the node attributes of a graph into forces and the edge attributes into springs. The algorithm finds an equilibrium position when the forces of the nodes are balanced by the forces on the springs. A full description of the algorithm is given in "The spring bounces back: Introduction to Strain Elevation Tension Spring embedding for network representation" (Bourne 2020). There is a website for the package providing documentation and vignettes at https://jonnob.github.io/rSETSe/index.html . This is a very niche package so please feel free to reach out to me on twitter or through email with questions.

Installation instructions

The package is available on CRAN and can be installed by running install.packages("rsetse").Alternatively it can be installed from github using the below method.

1. Open R/Rstudio and ensure that devtools has been installed
2. Run the following code library(devtools); install_github("JonnoB/rSETSe")
3. Load the package normally using library(rsetse)
4. All functions have help files e.g ?setse_auto

The package can also be downloaded or cloned then installed locally using the install function from devtools.

Basic use

library(rSETSe)

#prepares a graph for embedding using SETSe
set.seed(234) #set the random see for generating the network
g <- generate_peels_network(type = "E") %>%
prepare_edges(k = 500, distance = 1) %>%
#prepare the network for a binary embedding
prepare_categorical_force(., node_names = "name",
                     force_var = "class") 
                       
#Embedds using the bi-connected auto-parametrization algorithm.
#This method is strongly reccomended, it tends to be much faster and almost always converges
embeddings <- setse_bicomp(g,
                           force = "class_A",
                           tol = sum(abs(vertex_attr(g, "class_A")))/1000,
                           hyper_tol = 0.1,
                           hyper_iters = 3000,
                           verbose = T)

Cite

To cite rsetse in publications use: Bourne, J. The spring bounces back: introducing the strain elevation tension spring embedding algorithm for network representation. Appl Netw Sci 5, 88 (2020). https://doi.org/10.1007/s41109-020-00329-4