Optimizing Classification by Minimizing Table Deviance
opttdev is a iterative re-assignment classification algorithm that assigns samples to clusters to minimize the total deviance of a table with respect to the row-wise relative abundance of the elements
opttdev(comm,clustering,maxitr=100,minsiz=5)
comm |
a vegetation or other taxon table with samples as rows and taxa as columns |
clustering |
an index of cluster membership for each sample. May be either a numeric vector of length equal to the number of samples, or an object that inherits from class ‘cluster’ |
maxitr |
the maximum number of iterations to attempt |
minsiz |
the minimum size cluster to consider reassigning a sample out of |
Iterative re-allocation algorithms temporarily re-assign each sample to each
of the other possible clusters and calculate a goodness-of-clustering statistic for each
re-assignment. The best of all possible re-assignments is then executed and the algorithm
iterates until there are no more good re-assignments or the maximum number of iterations
is reached. In opttdev, the goodness-of-clustering statistic is total table deviance
as calculated by tabdev. See the help file for
tabdev for more detail.
a list which inherits from class ‘opttdev’, ‘clustering’ with components:
numitr |
the number of iterations performed |
dev |
a vector of total table deviance at each iteration of length ‘numitr’ |
clustering |
the vector of cluster memberships (as integers) for each sample |
Like many iterative re-assignment algorithms, opttdev is likely to be VERY slow from a random start or poor initial condition. opttdev is maybe better used to polish existing classifications
David W. Roberts droberts@montana.edu
## Not run: data(shoshveg) # returns a data.frame of vegetation
## Not run: data(shoshsite)
## Not run: res <- opttdev(shoshveg,
as.numeric(cut(shoshsite$elevation,5)))
## End(Not run)
## Not run: # likely to be VERY slowPlease choose more modern alternatives, such as Google Chrome or Mozilla Firefox.