我记得最普通的ERGM，被解释的网络只能是0-1
你的解释变量难道不应该也是26x26?

inf 估计是解释矩阵被认为为0


我曾经做过的

# adj_iobin  0-1 邻接方阵

io <- graph_from_adjacency_matrix(adj_iobin, mode = 'undirected', weighted = TRUE, add.colnames = 'label', diag = FALSE)

a_network <- network(adj_iobin, directed = FALSE)  #被解释网络


net_full <- asNetwork(make_full_graph(length(V(io)), directed = FALSE, loops = FALSE))
#解释网络  全连接

set.edge.value(net_full, "dist_io", as.matrix(distMatrix_io))
# distMatrix_io 是方阵，node--node距离，赋予到名为“dist_io"的属性

DSmod0 <- ergm(a_network ~ edges +  edgecov(net_full, "dist_io"), control = control.ergm(seed = 40))

我记得最普通的ERGM，被解释的网络只能是0-1
你的解释变量难道不应该也是26x26?

inf 估计是解释矩阵被认为为0


我曾经做过的

# adj_iobin  0-1 邻接方阵

io <- graph_from_adjacency_matrix(adj_iobin, mode = 'undirected', weighted = TRUE, add.colnames = 'label', diag = FALSE)

a_network <- network(adj_iobin, directed = FALSE)  #被解释网络


net_full <- asNetwork(make_full_graph(length(V(io)), directed = FALSE, loops = FALSE))
#解释网络  全连接

set.edge.value(net_full, "dist_io", as.matrix(distMatrix_io))
# distMatrix_io 是方阵，node--node距离，赋予到名为“dist_io"的属性

DSmod0 <- ergm(a_network ~ edges +  edgecov(net_full, "dist_io"), control = control.ergm(seed = 40))

应该把数据处理成26*26的处理成网络数据，然后26*9的作为节点数据加进网络属性，然后根据26*9的数据也可以计算边的属性，之后进行ergm就很简单了

看到一篇文章提到的是计数ergm，那被解释变量是不是就不要求必须是0或1

你好，net_full <- asNetwork(make_full_graph(length(V(io)), directed = FALSE, loops = FALSE)) 提示没有asNetwork这个函数，请问是怎么回事