使用 igraph 在二分网络上工作:基本度量问题(密度、归一化程度)
Working on bipartite networks with igraph : problem with basic measures (density, normalized degree)
我是二分网络分析的新手,我在基本测量方面遇到了一些麻烦。
我正在尝试在二分网络上工作 而不在 1 模式图中投影 。
我的问题来自这样一个事实,即 igraph 包允许创建二分图,但这些措施似乎并不适应这些图的特殊性。
所以,我的一般问题是当你直接在二分网络上工作时你会怎么做?
这里是密度的具体例子
## Working with an incidence matrix (sample) with 47 columns and 10 rows (unweighted / undirected)
# Want to compute basic global index like density with igraph
library(igraph)
g <- graph.incidence(m, directed = F )
graph.density(g) # result = 0.04636591
# Now trying to compute basic density for a bipartite graph without igraph (number of edges divided by the product of the two types of vertices)
library(Matrix)
d <- nnzero(m)/ (ncol(m)*nrow(m)) # result 0.1574468
# It seems that bipartite package does the job
library(bipartite)
networklevel(m, index=c("connectance")) # result 0.1574468
但是 bipartite 包非常针对生态学领域,很多措施都是针对食物网和物种间的相互作用而设计的(有些,比如聚类系数,似乎没有考虑到图的二分性质:例如计算4个周期)。
那么,有没有更简单的方法可以使用 igraph 在二分网络上工作?要测量一些全局指标(密度,4 周期的聚类系数,我知道 tnet 会这样做,但我的实际网络太大),并标准化局部指标,如度数、接近度、考虑二分特异性的介数中心性(如 Borgatti S.P., Everett M.G., 1997, « 2-mode data 的网络分析», Social Networks)?
如有任何建议,我们将不胜感激!
下面的代码重现了我的矩阵示例"m"
m <- structure(c(1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1,
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1,
0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0,
1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0,
0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0,
0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0), .Dim = c(10L, 47L), .Dimnames = list(
c("02723", "13963", "F3238", "02194", "15051", "04477", "02164",
"06283", "04080", "08304"), c("1185241", "170063", "10350868",
"217831", "2210247", "2262963", "1816670", "1848354", "2232593",
"146214", "1880252", "2261639", "2262581", "2158177", "1850147",
"2262912", "146412", "2262957", "1566083", "1841811", "146384",
"216281", "2220957", "1846986", "1951567", "1581130", "105343",
"1580240", "170654", "1796236", "1835553", "1835848", "146400",
"1174872", "1283240", "2253354", "1283617", "146617", "160263",
"2263115", "184745", "1809858", "1496747", "10346824", "148730",
"2262582", "146268")))
密度:你已经知道了
学位
degv1 <- degree(g, V(g)[type == FALSE])
degv2 <- degree(g, V(g)[type == TRUE])
归一化度数:除以其他节点类别的vcount
degnormv1 <- degv1/length(V(g)[type == TRUE])
degnormv2 <- degv2/length(V(g)[type == FALSE])
没有关于亲密度、介数和聚类系数的答案
对于归一化程度,这里是没有igraph的解决方案
normalizedegreeV1 <- data.frame(ND = colSums(m)/nrow(m))
normalizedegreeV2 <- data.frame(ND = rowSums(m)/ncol(m))
但这留下了关于中心性措施的其他问题悬而未决...
我是二分网络分析的新手,我在基本测量方面遇到了一些麻烦。 我正在尝试在二分网络上工作 而不在 1 模式图中投影 。 我的问题来自这样一个事实,即 igraph 包允许创建二分图,但这些措施似乎并不适应这些图的特殊性。
所以,我的一般问题是当你直接在二分网络上工作时你会怎么做?
这里是密度的具体例子
## Working with an incidence matrix (sample) with 47 columns and 10 rows (unweighted / undirected)
# Want to compute basic global index like density with igraph
library(igraph)
g <- graph.incidence(m, directed = F )
graph.density(g) # result = 0.04636591
# Now trying to compute basic density for a bipartite graph without igraph (number of edges divided by the product of the two types of vertices)
library(Matrix)
d <- nnzero(m)/ (ncol(m)*nrow(m)) # result 0.1574468
# It seems that bipartite package does the job
library(bipartite)
networklevel(m, index=c("connectance")) # result 0.1574468
但是 bipartite 包非常针对生态学领域,很多措施都是针对食物网和物种间的相互作用而设计的(有些,比如聚类系数,似乎没有考虑到图的二分性质:例如计算4个周期)。
那么,有没有更简单的方法可以使用 igraph 在二分网络上工作?要测量一些全局指标(密度,4 周期的聚类系数,我知道 tnet 会这样做,但我的实际网络太大),并标准化局部指标,如度数、接近度、考虑二分特异性的介数中心性(如 Borgatti S.P., Everett M.G., 1997, « 2-mode data 的网络分析», Social Networks)?
如有任何建议,我们将不胜感激!
下面的代码重现了我的矩阵示例"m"
m <- structure(c(1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1,
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1,
0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0,
1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0,
0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0,
0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 0, 0, 1, 0, 1, 0, 1, 0, 0, 1, 0), .Dim = c(10L, 47L), .Dimnames = list(
c("02723", "13963", "F3238", "02194", "15051", "04477", "02164",
"06283", "04080", "08304"), c("1185241", "170063", "10350868",
"217831", "2210247", "2262963", "1816670", "1848354", "2232593",
"146214", "1880252", "2261639", "2262581", "2158177", "1850147",
"2262912", "146412", "2262957", "1566083", "1841811", "146384",
"216281", "2220957", "1846986", "1951567", "1581130", "105343",
"1580240", "170654", "1796236", "1835553", "1835848", "146400",
"1174872", "1283240", "2253354", "1283617", "146617", "160263",
"2263115", "184745", "1809858", "1496747", "10346824", "148730",
"2262582", "146268")))
密度:你已经知道了
学位 degv1 <- degree(g, V(g)[type == FALSE]) degv2 <- degree(g, V(g)[type == TRUE])
归一化度数:除以其他节点类别的vcount
degnormv1 <- degv1/length(V(g)[type == TRUE]) degnormv2 <- degv2/length(V(g)[type == FALSE])
没有关于亲密度、介数和聚类系数的答案
对于归一化程度,这里是没有igraph的解决方案
normalizedegreeV1 <- data.frame(ND = colSums(m)/nrow(m))
normalizedegreeV2 <- data.frame(ND = rowSums(m)/ncol(m))
但这留下了关于中心性措施的其他问题悬而未决...