Mixture of Diagonal Gaussians is a subclass of Mixtureof Gaussians which
only uses diagonal of the covariance matrices in its fitness compuations.
Methods
|
|
__init__
evaluateFitness
|
|
__init__
|
__init__ (
self,
k,
means,
covariances,
weights=None,
)
Creates a new Mixture of Diagonal Gaussians (MoDG) Model, containing k
Gaussian clusters in d-dimensional space.
|
|
evaluateFitness
|
evaluateFitness ( self, data )
Return the fitness of the model given a paricular set of data. Using only
the diagonal covariances, if that fails, then the log-likelihood is set to
-MAX_FLOAT =~ -1e38.
|
|