Daniel@0: Daniel@0:
Daniel@0:Daniel@0: Daniel@0: mix = gmminit(mix, x, options) Daniel@0:Daniel@0: Daniel@0: Daniel@0:
mix = gmminit(mix, x, options)
uses a dataset x
Daniel@0: to initialise the parameters of a Gaussian mixture
Daniel@0: model defined by the data structure mix
. The k-means algorithm
Daniel@0: is used to determine the centres. The priors are computed from the
Daniel@0: proportion of examples belonging to each cluster.
Daniel@0: The covariance matrices are calculated as the sample covariance of the
Daniel@0: points associated with (i.e. closest to) the corresponding centres.
Daniel@0: For a mixture of PPCA model, the PPCA decomposition is calculated
Daniel@0: for the points closest to a given centre.
Daniel@0: This initialisation can be used as the starting point for training the
Daniel@0: model using the EM algorithm.
Daniel@0:
Daniel@0: Daniel@0: Daniel@0: mix = gmm(3, 2); Daniel@0: options = foptions; Daniel@0: options(14) = 5; Daniel@0: mix = gmminit(mix, data, options); Daniel@0:Daniel@0: Daniel@0: This code sets up a Gaussian mixture model with 3 centres in 2 dimensions, and Daniel@0: then initialises the parameters from the data set
data
with 5 iterations
Daniel@0: of the k means algorithm.
Daniel@0:
Daniel@0: gmm
Copyright (c) Ian T Nabney (1996-9) Daniel@0: Daniel@0: Daniel@0: Daniel@0: