Determine highly correlated variables

This function searches through a correlation matrix and returns a vector of integers corresponding to columns to remove to reduce pair-wise correlations.

findCorrelation(x, cutoff = 0.9, verbose = FALSE, names = FALSE,
  exact = ncol(x) < 100)

Arguments

x	A correlation matrix
cutoff	A numeric value for the pair-wise absolute correlation cutoff
verbose	A boolean for printing the details
names	a logical; should the column names be returned (TRUE) or the column index (FALSE)?
exact	a logical; should the average correlations be recomputed at each step? See Details below.

Value

A vector of indices denoting the columns to remove (when names = TRUE) otherwise a vector of column names. If no correlations meet the criteria, integer(0) is returned.

Details

The absolute values of pair-wise correlations are considered. If two variables have a high correlation, the function looks at the mean absolute correlation of each variable and removes the variable with the largest mean absolute correlation.

Using exact = TRUE will cause the function to re-evaluate the average correlations at each step while exact = FALSE uses all the correlations regardless of whether they have been eliminated or not. The exact calculations will remove a smaller number of predictors but can be much slower when the problem dimensions are "big".

There are several function in the subselect package (leaps, genetic, anneal) that can also be used to accomplish the same goal but tend to retain more predictors.

See also

leaps, genetic, anneal, findLinearCombos

Examples

R1 <- structure(c(1, 0.86, 0.56, 0.32, 0.85, 0.86, 1, 0.01, 0.74, 0.32,
                  0.56, 0.01, 1, 0.65, 0.91, 0.32, 0.74, 0.65, 1, 0.36,
                  0.85, 0.32, 0.91, 0.36, 1),
                .Dim = c(5L, 5L))
colnames(R1) <- rownames(R1) <- paste0("x", 1:ncol(R1))
R1
#>      x1   x2   x3   x4   x5
#> x1 1.00 0.86 0.56 0.32 0.85
#> x2 0.86 1.00 0.01 0.74 0.32
#> x3 0.56 0.01 1.00 0.65 0.91
#> x4 0.32 0.74 0.65 1.00 0.36
#> x5 0.85 0.32 0.91 0.36 1.00

findCorrelation(R1, cutoff = .6, exact = FALSE)
#> [1] 4 5 1 3
findCorrelation(R1, cutoff = .6, exact = TRUE)
#> [1] 1 5 4
findCorrelation(R1, cutoff = .6, exact = TRUE, names = FALSE)
#> [1] 1 5 4


R2 <- diag(rep(1, 5))
R2[2, 3] <- R2[3, 2] <- .7
R2[5, 3] <- R2[3, 5] <- -.7
R2[4, 1] <- R2[1, 4] <- -.67

corrDF <- expand.grid(row = 1:5, col = 1:5)
corrDF$correlation <- as.vector(R2)
levelplot(correlation ~ row + col, corrDF)

findCorrelation(R2, cutoff = .65, verbose = TRUE)
#> Compare row 3  and column  2 with corr  0.7 
#>   Means:  0.35 vs 0.215 so flagging column 3 
#> Compare row 1  and column  4 with corr  0.67 
#>   Means:  0.223 vs 0.074 so flagging column 1 
#> [1] 3 1

findCorrelation(R2, cutoff = .99, verbose = TRUE)
#> All correlations <= 0.99 
#> integer(0)