Reconstruct a symmetric matrix from a distance (lower-triangular) representation
Assuming the input vector contains a vectorized form of the distance representation of a symmetric matrix, this function creates the corresponding matrix. This is useful when re-forming symmetric matrices that have been vectorized to save storage space.
lowerTri2matrix(x, diag = 1)
x |
a numeric vector |
diag |
value to be put on the diagonal. Recycled if necessary. |
The function assumes that x contains the vectorized form of the distance representation of a
symmetric matrix. In particular, x must have a length that can be expressed as n*(n-1)/2, with n an
integer. The result of the function is then an n times n matrix.
A symmetric matrix whose lower triangle is given by x.
Peter Langfelder
# Create a symmetric matrix m = matrix(c(1:16), 4,4) mat = (m + t(m)); diag(mat) = 0; # Print the matrix mat # Take the lower triangle and vectorize it (in two ways) x1 = mat[lower.tri(mat)] x2 = as.vector(as.dist(mat)) all.equal(x1, x2) # The vectors are equal # Turn the vectors back into matrices new.mat = lowerTri2matrix(x1, diag = 0); # Did we get back the same matrix? all.equal(mat, new.mat)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.