| Title: | rMultiNet: An R Package For Multilayer Networks Analysis |
|---|---|
| Description: | https://github.com/ChenyuzZZ73/rMultiNet |
| Authors: | Chenyu Ren, Ting Li, Zhongyuan Lyu, Dong Xia |
| Maintainer: | <[email protected]> |
| License: | MIT |
| Version: | 0.1.0 |
| Built: | 2026-10-11 15:21:13 UTC |
| Source: | https://github.com/chenyuzzz73/rmultinet |
a method to visuilize the result of embedding
Community_cluster_km(embedding, type, cluster_number)Community_cluster_km(embedding, type, cluster_number)
embedding |
the embedding result of layers or nodes |
type |
N for layer n for nodes |
cluster_number |
the number of cluster |
a method to visuilize the result of embedding
Embedding_network(network_membership, L, paxis = 2)Embedding_network(network_membership, L, paxis = 2)
network_membership |
|
L |
the number of network membership(n or L) |
paxis |
= 2 the number of axis to plot |
a method to generate a tensor for multilayer network
logi_fun(x, scale_par)logi_fun(x, scale_par)
scale_par |
the scale parameter(default as 1) |
n |
number of nodes |
M |
types of networks |
d |
rank of U matrix |
L |
number of layers |
U_mean |
the mean of U with the Norm distribution, sd = 1 |
cmax |
the upper bound of C tensor |
int_type |
way of generating C tensor included "Uniform" and "Norm" |
kernel_fun |
way of generating the adjacency tensor included "logit" and "probit" |
Pi_network_type |
size of network types (default as 1/m, an m-dimensional vector) |
a method to generate a tensor for multilayer network
generate_sbm(n, K, r, d, membership)generate_sbm(n, K, r, d, membership)
n |
the number of nodes |
K |
number of communities |
r |
out-in ratio (q/p, a L-dimensional vector) |
d |
average degree (a L-dimensional vector) |
m |
types of networks |
L |
number of layers |
A function for initialization
Init_UW(tensor_A, n, r, M, Utrue, Wtrue, pertub = 0.1, int_type)Init_UW(tensor_A, n, r, M, Utrue, Wtrue, pertub = 0.1, int_type)
n |
nodes |
r |
rank of U |
M |
the number of network types |
pertub |
the upper bound of Uniform distribution |
int_type |
way of initialize UW included "spec", "rand" and "warm" |
gen_list |
adj_tensor, Theta |
list(adj_tensor,U_int, W_int, delta)
A function for initialization
InitializationMMSBM(tnsr, ranks = c(2, 2, 2))InitializationMMSBM(tnsr, ranks = c(2, 2, 2))
tnsr |
the tensor of network |
ranks=NULL |
the rank of low_rank tensor |
A function for tensor decomposition
norm_vec(x)norm_vec(x)
tnsr |
adjacency tensor (n*n*L) |
ranks |
rank of core tensor |
U_0_list |
initialization |
delta1 |
reguliazation for mode1 and mode2 |
delta2 |
reguliazation for mode3 |
max_iter |
max times of iteration |
tol |
covergence tolerance |
A function for tensor decomposition
reg_fun(U, delta)reg_fun(U, delta)
Ini_list |
list(adj_tensor,U_int, W_int, delta) delta for incoherent parameters: (delta_1,delta_2,delta_3) |
Cmax |
upper bound |
eta_outer |
learning rate eta |
tmax_outer |
number of iterations in gradient descent |
p_type |
type of link function included "logit", "probit" and "poisson" |
rd |
if use stochastic sampling ("rand" or "Non") |
scale_par |
scale parameter (sigma in the paper) |
sample_size |
A function for tensor decomposition
SpecClustering(tnsr, rank, embedding_type = "Layer")SpecClustering(tnsr, rank, embedding_type = "Layer")
matrx |
an n by n matrix |
ncol |
number of columns of the output matrix U |
U n by rank numeric matrix that contains the rank tops eigenvectors of Laplacian matrix as column