Mixture

class mixture.Mixture(mixtureFileName='', mixtureFileFormat='pickle', name='empty')

A class for Gaussian Mixture Model storage. For more details about Gaussian Mixture Models (GMM) you can refer to [Bimbot04].

Attr w:array of weight parameters Attr mu:ndarray of mean parameters, each line is one distribution Attr invcov:ndarray of inverse co-variance parameters, 2-dimensional for diagonal co-variance distribution 3-dimensional for full co-variance Attr invchol:3-dimensional ndarray containing lower cholesky decomposition of the inverse co-variance matrices Attr cst:array of constant computed for each distribution Attr det:array of determinant for each distribution

EM_split(cep, distrib_nb, iterations=[1, 2, 2, 4, 4, 4, 4, 8, 8, 8, 8, 8, 8], numThread=1, llk_gain=0.01)

Expectation-Maximization estimation of the Mixture parameters.

Parameters:

• cep – set of feature frames to consider
• distrib_nb – final number of distributions to reach
• iterations – a list of number of iterations to perform before spliting the distributions.
• numThread – number of thread to launch for parallel computing

Return llk:

a list of log-likelihoods obtained after each iteration

EM_uniform(cep, distribNb, iteration_min=3, iteration_max=10, llk_gain=0.01, do_init=True)

Expectation-Maximization estimation of the Mixture parameters.

Parameters:

• cep – set of feature frames to consider
• distribNb – number of distributions
• iteration – number of iterations to perform.
• numThread – number of thread to launch for parallel computing

Return llk:

a list of log-likelihoods obtained after each iteration

compute_log_posterior_probabilities(cep, mu=None)

Compute log posterior probabilities for a set of feature frames.

Parameters:

• cep – a set of feature frames in a ndarray, one feature per row
• mu – a mean super-vector to replace the ubm’s one. If it is an empty vector, use the UBM

Returns:

A ndarray of log-posterior probabilities corresponding to the input feature set.

dim()

Return the dimension of distributions of the Mixture

Returns:an integer, size of the acoustic vectors

distrib_nb()

Return the number of distribution of the Mixture

Returns:the number of distribution in the Mixture

get_invcov_super_vector()

Return Inverse covariance super-vector

Returns:an array, super-vector of the inverse co-variance coefficients

get_mean_super_vector()

Return mean super-vector

Returns:an array, super-vector of the mean coefficients

read_alize(mixtureFileName)

Read a Mixture in alize raw format

Parameters:mixtureFileName – name of the file to read from

read_htk(mixtureFileName, beginHmm=False, state2=False)

Read a Mixture in HTK format

Parameters:

• mixtureFileName – name of the file to read from
• beginHmm – boolean
• state2 – boolean

read_pickle(input_filename)

Read IdMap in PICKLE format.

Parameters:inputFileName – name of the file to read from

save_alize(mixtureFileName)

Save a mixture in alize raw format

Parameters:mixtureFileName – name of the file to write in

save_htk(mixtureFileName)

Save a Mixture in HTK format

Parameters:mixtureFileName – the name of the file to write in

save_pickle(outputFileName)

Save Ndx in PICKLE format. Convert all data into float32 before saving, note that the conversion doesn’t apply in Python 2.X

Parameters:outputFilename – name of the file to write to

sv_size()

Return the dimension of the super-vector

Returns:an integer, size of the mean super-vector

validate()

Verify the format of the Mixture

Returns:a boolean giving the status of the Mixture

varianceControl(cov, flooring, ceiling, cov_ctl)

varianceControl for Mixture (florring and ceiling)

Parameters:

• flooring – float, florring value
• ceiling – float, ceiling value
• covSignal – co-variance to consider for flooring and ceiling