将自定义 2D 稀疏矩阵转换为网络

Question

我有这样一个从程序中导出的数组，看起来像：

1 2:1.827411e-02 3:5.355330e-02 4:1.827411e-02 5:1.827411e-02
2 1:1.827411e-02 3:1.903553e-02 4:4.568528e-03 5:4.568528e-03
3 1:5.355330e-02 2:1.903553e-02 4:1.903553e-02 5:1.903553e-02 6:7.461929e-02 11:3.350254e-02
4 1:1.827411e-02 2:4.568528e-03 3:1.903553e-02 5:4.568528e-03
5 1:1.827411e-02 2:4.568528e-03 3:1.903553e-02 4:4.568528e-03
6 3:7.461929e-02 7:1.903553e-02 8:1.903553e-02 9:5.355330e-02 10:1.903553e-02 11:3.350254e-02
7 6:1.903553e-02 8:4.568528e-03 9:1.827411e-02 10:4.568528e-03
8 6:1.903553e-02 7:4.568528e-03 9:1.827411e-02 10:4.568528e-03
9 6:5.355330e-02 7:1.827411e-02 8:1.827411e-02 10:1.827411e-02
10 6:1.903553e-02 7:4.568528e-03 8:4.568528e-03 9:1.827411e-02
11 3:3.350254e-02 6:3.350254e-02

每行描述行的第一个数字和冒号之前的数字之间的边的权重（冒号之后）。

即第一行：

the weight between 1 and 2 is 1.827411e-02
the weight between 1 and 3 is 5.355330e-02
the weight between 1 and 4 is 1.827411e-02
the weight between 1 and 5 is 1.827411e-02

其余行以此类推。

这个信息我想以某种方式将其转换为 igraph 或网络元素，以便对其进行更多分析。有没有有效的方法来做到这一点？

Answer 1

您可以使用一点正则表达式来提取相关值以形成加权边缘列表。

# parent node : grab the first number
parent <- sub("(\\d+ ).*$", "\\1", r) 

# child node: grab everything after the first number and split it
child <- strsplit(sub("\\d+ (.*$)", "\\1", r), ":| ")

# cbind the parent node to the child    
dat2 <- cbind(rep(parent, lengths(child)/2), matrix(unlist(child), nc=2, byrow=TRUE))

mode(dat2) = "numeric" # change to numeric


# read in a graph: the weights are in the edge attributes
g <- igraph::graph_from_data_frame(dat2)

---

数据

假设您的数据在文本文件中采用这种形式（将 textConnection(txt) 更改为文件路径）

txt <- 
  '1 2:1.827411e-02 3:5.355330e-02 4:1.827411e-02 5:1.827411e-02
2 1:1.827411e-02 3:1.903553e-02 4:4.568528e-03 5:4.568528e-03
3 1:5.355330e-02 2:1.903553e-02 4:1.903553e-02 5:1.903553e-02 6:7.461929e-02 11:3.350254e-02
4 1:1.827411e-02 2:4.568528e-03 3:1.903553e-02 5:4.568528e-03
5 1:1.827411e-02 2:4.568528e-03 3:1.903553e-02 4:4.568528e-03
6 3:7.461929e-02 7:1.903553e-02 8:1.903553e-02 9:5.355330e-02 10:1.903553e-02 11:3.350254e-02
7 6:1.903553e-02 8:4.568528e-03 9:1.827411e-02 10:4.568528e-03
8 6:1.903553e-02 7:4.568528e-03 9:1.827411e-02 10:4.568528e-03
9 6:5.355330e-02 7:1.827411e-02 8:1.827411e-02 10:1.827411e-02
10 6:1.903553e-02 7:4.568528e-03 8:4.568528e-03 9:1.827411e-02
11 3:3.350254e-02 6:3.350254e-02'

r <- readLines(textConnection(txt))