CODEWITHSUNDEEP
r
Shaun
Shaun

Reputation: 949

Replace the values in the diagonal of a (correlation) matrix with zero, but keep all other 1 values in the matrix

I have a correlation matrix, and using it to identify where variables are exactly correlated with other variables, knowing a correlation=1 is exact match. To better explain, here is some simulated data, to create some variables which are identical in values.

set.seed(123)
df <- data.frame(replicate(11,sample(1:1000,10,rep=TRUE)))
df <- cbind(df,df[,1:3],df[,1] )
names(df) <- c(paste0("X", 1:15))
cor_matrix <- round(cor(df, use="complete.obs", method="pearson"),2)
cor_matrix
> cor_matrix
       X1    X2    X3    X4    X5    X6    X7    X8    X9   X10   X11   X12   X13   X14   X15
X1   1.00  0.08 -0.69  0.10 -0.35 -0.39  0.26 -0.17  0.06 -0.03  0.12  1.00  0.08 -0.69  1.00
X2   0.08  1.00 -0.11  0.15 -0.30 -0.41  0.05  0.09 -0.70  0.67 -0.10  0.08  1.00 -0.11  0.08
X3  -0.69 -0.11  1.00 -0.18 -0.05  0.29 -0.42 -0.10 -0.01 -0.12  0.03 -0.69 -0.11  1.00 -0.69
X4   0.10  0.15 -0.18  1.00  0.24  0.37  0.31  0.18  0.03 -0.09  0.28  0.10  0.15 -0.18  0.10
X5  -0.35 -0.30 -0.05  0.24  1.00  0.33  0.36  0.27  0.34 -0.08 -0.21 -0.35 -0.30 -0.05 -0.35
X6  -0.39 -0.41  0.29  0.37  0.33  1.00 -0.55  0.11  0.55 -0.78  0.15 -0.39 -0.41  0.29 -0.39
X7   0.26  0.05 -0.42  0.31  0.36 -0.55  1.00  0.01 -0.25  0.54  0.10  0.26  0.05 -0.42  0.26
X8  -0.17  0.09 -0.10  0.18  0.27  0.11  0.01  1.00  0.06  0.02 -0.47 -0.17  0.09 -0.10 -0.17
X9   0.06 -0.70 -0.01  0.03  0.34  0.55 -0.25  0.06  1.00 -0.82 -0.26  0.06 -0.70 -0.01  0.06
X10 -0.03  0.67 -0.12 -0.09 -0.08 -0.78  0.54  0.02 -0.82  1.00 -0.17 -0.03  0.67 -0.12 -0.03
X11  0.12 -0.10  0.03  0.28 -0.21  0.15  0.10 -0.47 -0.26 -0.17  1.00  0.12 -0.10  0.03  0.12
X12  1.00  0.08 -0.69  0.10 -0.35 -0.39  0.26 -0.17  0.06 -0.03  0.12  1.00  0.08 -0.69  1.00
X13  0.08  1.00 -0.11  0.15 -0.30 -0.41  0.05  0.09 -0.70  0.67 -0.10  0.08  1.00 -0.11  0.08
X14 -0.69 -0.11  1.00 -0.18 -0.05  0.29 -0.42 -0.10 -0.01 -0.12  0.03 -0.69 -0.11  1.00 -0.69
X15  1.00  0.08 -0.69  0.10 -0.35 -0.39  0.26 -0.17  0.06 -0.03  0.12  1.00  0.08 -0.69  1.00

As you can see correlation value of 1 can be found X1 with X12 and X15. X2 and X13, X3 AND X14. What I want to do is remove all values in the matrix that are not 1, (replace the unwanted values with zero) but will need to remove the 1s that are on the diagonal, where it's directly related to itself. Any help or ideas would be greatly appreciated. Many thanks

Upvotes: 2

Views: 1183

Answers (1)

akrun
akrun

Reputation: 887118

We can assign the diag elements to 0

diag(cor_matrix) <- 0

-output

cor_matrix
       X1    X2    X3    X4    X5    X6    X7    X8    X9   X10   X11   X12   X13   X14   X15
X1   0.00  0.08 -0.69  0.10 -0.35 -0.39  0.26 -0.17  0.06 -0.03  0.12  1.00  0.08 -0.69  1.00
X2   0.08  0.00 -0.11  0.15 -0.30 -0.41  0.05  0.09 -0.70  0.67 -0.10  0.08  1.00 -0.11  0.08
X3  -0.69 -0.11  0.00 -0.18 -0.05  0.29 -0.42 -0.10 -0.01 -0.12  0.03 -0.69 -0.11  1.00 -0.69
X4   0.10  0.15 -0.18  0.00  0.24  0.37  0.31  0.18  0.03 -0.09  0.28  0.10  0.15 -0.18  0.10
X5  -0.35 -0.30 -0.05  0.24  0.00  0.33  0.36  0.27  0.34 -0.08 -0.21 -0.35 -0.30 -0.05 -0.35
X6  -0.39 -0.41  0.29  0.37  0.33  0.00 -0.55  0.11  0.55 -0.78  0.15 -0.39 -0.41  0.29 -0.39
X7   0.26  0.05 -0.42  0.31  0.36 -0.55  0.00  0.01 -0.25  0.54  0.10  0.26  0.05 -0.42  0.26
X8  -0.17  0.09 -0.10  0.18  0.27  0.11  0.01  0.00  0.06  0.02 -0.47 -0.17  0.09 -0.10 -0.17
X9   0.06 -0.70 -0.01  0.03  0.34  0.55 -0.25  0.06  0.00 -0.82 -0.26  0.06 -0.70 -0.01  0.06
X10 -0.03  0.67 -0.12 -0.09 -0.08 -0.78  0.54  0.02 -0.82  0.00 -0.17 -0.03  0.67 -0.12 -0.03
X11  0.12 -0.10  0.03  0.28 -0.21  0.15  0.10 -0.47 -0.26 -0.17  0.00  0.12 -0.10  0.03  0.12
X12  1.00  0.08 -0.69  0.10 -0.35 -0.39  0.26 -0.17  0.06 -0.03  0.12  0.00  0.08 -0.69  1.00
X13  0.08  1.00 -0.11  0.15 -0.30 -0.41  0.05  0.09 -0.70  0.67 -0.10  0.08  0.00 -0.11  0.08
X14 -0.69 -0.11  1.00 -0.18 -0.05  0.29 -0.42 -0.10 -0.01 -0.12  0.03 -0.69 -0.11  0.00 -0.69
X15  1.00  0.08 -0.69  0.10 -0.35 -0.39  0.26 -0.17  0.06 -0.03  0.12  1.00  0.08 -0.69  0.00

Upvotes: 3

Related Questions