cluster_edge_betweenness 在来自命名 one-mode 矩阵的 igraph 中

Question

我想使用 igraph 函数 cluster_edge_betweenness 聚类一个 one-mode 网络矩阵进行绘图，据我所知，这是可行的：

library("igraph")

adjm <- matrix(sample(0:1, 100, replace=TRUE, prob=c(0.9,0.1)), nc=10)
colnames(adjm) <- LETTERS[1:10]
row.names(adjm) <- LETTERS[1:10]

g <- graph_from_adjacency_matrix(adjm,
                                 mode = "undirected",
                                 weighted = "weight",
                                 diag = TRUE,
                                 add.colnames = NULL)
c <- cluster_edge_betweenness(g,
                              weights = igraph::E(g)$weight, 
                              edge.betweenness = TRUE, 
                              merges = TRUE, 
                              bridges = TRUE,
                              modularity = FALSE, 
                              membership = FALSE)

但是，这给了我错误：

Error in names(res) <- communities$names : 
'names' attribute [10] must be the same length as the vector [0]

这很奇怪，因为在 add.colnames = NULL 时应该分配名称，至少如果我正确理解帮助的话：

add.colnames
Character scalar, whether to add the column names as vertex attributes. If it is ‘NULL’ (the default) then, if present, column names are added as vertex attribute ‘name’.

然后，更奇怪的是，尽管有错误（不是警告！）cluster_edge_betweenness 实际上生成了我想要的对象并且我可以绘制它：

hc <- as.hclust(c, hang = -1, use.modularity = FALSE)
plot(hc)

这是怎么回事？我应该在他们的 GitHub 页面上提出问题还是我误解了什么？

注意：使用 igraph 1.2.1 测试。

Answer 1

函数中抛出错误membership:

function (communities) 
{
    if (!is.null(communities$membership)) {
        res <- communities$membership
    }
    else if (!is.null(communities$merges) && !is.null(communities$modularity)) {
        res <- community.to.membership2(communities$merges, communities$vcount, 
            which.max(communities$modularity))
    }
    else {
        stop("Cannot calculate community membership")
    }
    if (igraph_opt("add.vertex.names") && !is.null(communities$names)) {
        names(res) <- communities$names
    }
    class(res) <- "membership"
    res
}

问题似乎是 communities$membership 是 numeric(0)。因此 res 也变成了 numeric(0) 并且，因为

communities$names
# [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J"

我们在 names(res) <- communities$names 中得到了错误。

因此，它看起来像是来自某个地方的错误，正如我评论的那样，快速修复是在调用 cluster_edge_betweenness.

时设置 membership = TRUE

cluster_edge_betweenness 在来自命名 one-mode 矩阵的 igraph 中

cluster_edge_betweenness in igraph from named one-mode matrix

r

igraph

network-analysis