AUXIL\
gpuem.pro
top GPUEM
GPUEM, G, U, Ms, Ps, Cs [, unfrozen=unfrozen] [, wnd=wnd], maxiter=maxiter, miniter=miniter [, pdens=pdens] [, fhv=fhv] [, T0=T0] [, verbose=verbose], num_cols=num_cols, num_rows=num_rows [, beta=beta] [, status=status]
Clustering with Gaussian mixtures, EM algorithm:
Gath, I. and Geva, A. B. (1989). Unsupervised
optimal fuzzy clustering.
IEEE Transactions on Pattern Analysis and
Machine Intelligence, 3(3), 773–781.
Parameters
- G in required
observations data array (NNxn ) (row vectors)
- U in out required
initial class probability membership array (nxK) (column vectors)
- Ms out
cluster means matrix (row vectors)
- Ps out
cluster priors
- Cs out required
cluster covariance matrices (KxNNxNN)
Keywords
- unfrozen in optional
indices of the observations which take part in the iteration (default all)
- wnd in optional
window for displaying the log likelihood
- maxiter
- miniter
- pdens out optional
class partition densities
- fhv out optional
class fuzzy hypervolumes
- T0 in optional
initial annealing temperature (default 1.0)
- verbose in optional
set to print output info to IDL log (default 0)
- num_cols in required
if beta>0 image dimensions
- num_rows
- beta in optional
spatial field parameter
- status out optional
0 if successful else 1
Author information
- Author:
Mort Canty (2009)
Other attributes
- Uses:
COYOTE:, ,
, gpuconvol2D:GPULIB
File attributes
| Modification date: | Wed Jun 12 06:53:45 2013 |
| Lines: | 290 |
| Docformat: | rst rst |