hdist-class {hopach} | R Documentation |
Class hdist was created to take advantage of the structure innate to symmetric matrices. It stores only the lower triangle of the matrix, thus reducing the size (and memory usage) from n x n to [n x (n - 1)] / 2.
Like a matrix, a hdist object is subsettable; thus, hdist[i,j] will return the value at row 'i' column 'j'. Most valid indices for a matrix are also valid for a hdist object. (See examples below)
Data
:"numeric"
a vector containing the stacked columns of
the lower triangle of a symmetric matrix – often the symmetric matrix is a
distance matrix.Size
:"numeric"
the dimension of the
symmetric matrix, from which Data
was constructed.Labels
:"numeric"
a list of values of
length Size
to allow for pretty printing.Call
:"character"
a character string
specifying the method used to create the distance matrix
from which Data
was constructed.signature{Data = "numeric", Size = "numeric", Labels = "numeric", Call = "character"}
: Create a new hdist object. signature{from = "matrix"}
: Converts a matrix to a hdist object. signature(x = "hdist")
: Converts a hdist object to a matrix. signature(x = "hdist", mode = "missing")
: Returns the hdist object as a vector. signature(x = "hdist")
: Subsetting function for hdist objects. See examples and warning. signature(from = "matrix", to = "hdist")
: Converts a matrix to a hdist object. signature(from = "hdist", to = "matrix")
: Converts a hdist object to a matrix. signature(x = "hdist")
: Returns the dimension of the hdist object if expanded to a square matrix.signature(object = "hdist")
: Returns the labels used for printing. signature(x = "hdist")
: Returns the number of rows in hdist object. signature(object = "hdist")
: Prints the hdist object. A hdist object is NOT closed under the subsetting operation. For instance, if a 100 x 100 symmetric matrix is stored as an hdist object, hdist[c(3,4,5),c(7,8,9)] will return a 3 x 3 matrix, since the subsetting will not result in a symmetric matrix. However, if index i = j, then subsetting a hdist object will result in a symmetric matrix, and thus a hdist object will be returned. (See examples below)
Thank you to Larry Tai for his assistance creating run-time comparisons.
Katherine S. Pollard <kpollard@gladstone.ucsf.edu> and Gregory D. Wall <gwall@wald.ucdavis.edu>
van der Laan, M.J. and Pollard, K.S. A new algorithm for hybrid hierarchical clustering with visualization and the bootstrap. Journal of Statistical Planning and Inference, 2003, 117, pp. 275-303.
http://www.stat.berkeley.edu/~laan/Research/Research_subpages/Papers/hopach.pdf
http://www.bepress.com/ucbbiostat/paper107/
http://www.stat.berkeley.edu/~laan/Research/Research_subpages/Papers/jsmpaper.pdf
Kaufman, L. and Rousseeuw, P.J. (1990). Finding Groups in Data: An Introduction to Cluster Analysis. Wiley, New York.
showClass("hdist") library(hopach) X <- matrix(rnorm(60,mean=10,sd=2),nrow=10,ncol=6,byrow=TRUE) dmat <- disscosangle(X) dmat str(dmat) # Examples where subsetting a hdist object returns a matrix... dmat[c(3,4,5),c(5,6,7,8)] dmat[c(TRUE,FALSE),c(FALSE,TRUE)] dmat[c(4,5,6), ] # Examples where subsetting a hdist object returns a hdist object... dmat[c(3,4,5,6,7),c(3,4,5,6,7)] dmat[c(TRUE,FALSE),c(TRUE,FALSE)] # Expand hdist object to a symmetric matrix... as.matrix(dmat)