将相关矩阵转换为R中的Pajek输入文件

Question

I need to convert a correlation matrix to a Pajek input file for networks analysis. More specifically I'm trying to use R to convert a matrix of correlation p-values to a list of ”rowname columnname” for the significant correlations. That is a list of each of the significant correlations between quite a lot of variables. If I have variables a,b,c,d and a,c; b,d and a,d were correlated, I'd like a list as follows:

a b;
b d;
a d

So far my inadequate R skills have enabled me to generate a correlation p-value matrix, to insert NA on and below the diagonal (to avoid meaningless and duplicated correlations) and to substitute the p-values with FALSE/TRUE if the p-value is non-significant/significant. But now I am stuck and have not been able to google my way out of it.

Answer 1

Here's an example that may help with some basics:

#Create a matrix
m <- matrix(1:16,4,4)
rownames(m) <- letters[1:4]
colnames(m) <- letters[1:4]
m
  a b  c  d
a 1 5  9 13
b 2 6 10 14
c 3 7 11 15
d 4 8 12 16

#Identify the indices for entries in m
# that are greater than 10 
m1 <- which(m > 10, arr.ind = TRUE)

#Row and column names of those entries
# greater than 10. Notice the use of subsetting
# via [. 
cbind(rownames(m)[m1[,1]],colnames(m)[m1[,2]])
     [,1] [,2]
[1,] "c"  "c" 
[2,] "d"  "c" 
[3,] "a"  "d" 
[4,] "b"  "d" 
[5,] "c"  "d" 
[6,] "d"  "d" 

As with anything in R, there are many ways to do stuff like this, but this should give you some useful tools to work with.