MLPACK  1.0.7
em_fit.hpp
Go to the documentation of this file.
1 
23 #ifndef __MLPACK_METHODS_GMM_EM_FIT_HPP
24 #define __MLPACK_METHODS_GMM_EM_FIT_HPP
25 
26 #include <mlpack/core.hpp>
27 
28 // Default clustering mechanism.
30 // Default covariance matrix constraint.
32 
33 namespace mlpack {
34 namespace gmm {
35 
49 template<typename InitialClusteringType = kmeans::KMeans<>,
50  typename CovarianceConstraintPolicy = PositiveDefiniteConstraint>
51 class EMFit
52 {
53  public:
71  EMFit(const size_t maxIterations = 300,
72  const double tolerance = 1e-10,
73  InitialClusteringType clusterer = InitialClusteringType(),
74  CovarianceConstraintPolicy constraint = CovarianceConstraintPolicy());
75 
86  void Estimate(const arma::mat& observations,
87  std::vector<arma::vec>& means,
88  std::vector<arma::mat>& covariances,
89  arma::vec& weights);
90 
103  void Estimate(const arma::mat& observations,
104  const arma::vec& probabilities,
105  std::vector<arma::vec>& means,
106  std::vector<arma::mat>& covariances,
107  arma::vec& weights);
108 
110  const InitialClusteringType& Clusterer() const { return clusterer; }
112  InitialClusteringType& Clusterer() { return clusterer; }
113 
115  const CovarianceConstraintPolicy& Constraint() const { return constraint; }
117  CovarianceConstraintPolicy& Constraint() { return constraint; }
118 
120  size_t MaxIterations() const { return maxIterations; }
122  size_t& MaxIterations() { return maxIterations; }
123 
125  double Tolerance() const { return tolerance; }
127  double& Tolerance() { return tolerance; }
128 
129  private:
140  void InitialClustering(const arma::mat& observations,
141  std::vector<arma::vec>& means,
142  std::vector<arma::mat>& covariances,
143  arma::vec& weights);
144 
155  double LogLikelihood(const arma::mat& data,
156  const std::vector<arma::vec>& means,
157  const std::vector<arma::mat>& covariances,
158  const arma::vec& weights) const;
159 
163  double tolerance;
165  InitialClusteringType clusterer;
167  CovarianceConstraintPolicy constraint;
168 };
169 
170 }; // namespace gmm
171 }; // namespace mlpack
172 
173 // Include implementation.
174 #include "em_fit_impl.hpp"
175 
176 #endif
This class contains methods which can fit a GMM to observations using the EM algorithm.
Definition: em_fit.hpp:51
double & Tolerance()
Modify the tolerance for the convergence of the EM algorithm.
Definition: em_fit.hpp:127
const CovarianceConstraintPolicy & Constraint() const
Get the covariance constraint policy class.
Definition: em_fit.hpp:115
size_t maxIterations
Maximum iterations of EM algorithm.
Definition: em_fit.hpp:161
CovarianceConstraintPolicy constraint
Object which applies constraints to the covariance matrix.
Definition: em_fit.hpp:167
size_t & MaxIterations()
Modify the maximum number of iterations of the EM algorithm.
Definition: em_fit.hpp:122
size_t MaxIterations() const
Get the maximum number of iterations of the EM algorithm.
Definition: em_fit.hpp:120
InitialClusteringType & Clusterer()
Modify the clusterer.
Definition: em_fit.hpp:112
EMFit(const size_t maxIterations=300, const double tolerance=1e-10, InitialClusteringType clusterer=InitialClusteringType(), CovarianceConstraintPolicy constraint=CovarianceConstraintPolicy())
Construct the EMFit object, optionally passing an InitialClusteringType object (just in case it needs...
CovarianceConstraintPolicy & Constraint()
Modify the covariance constraint policy class.
Definition: em_fit.hpp:117
double LogLikelihood(const arma::mat &data, const std::vector< arma::vec > &means, const std::vector< arma::mat > &covariances, const arma::vec &weights) const
Calculate the log-likelihood of a model.
double Tolerance() const
Get the tolerance for the convergence of the EM algorithm.
Definition: em_fit.hpp:125
const InitialClusteringType & Clusterer() const
Get the clusterer.
Definition: em_fit.hpp:110
void InitialClustering(const arma::mat &observations, std::vector< arma::vec > &means, std::vector< arma::mat > &covariances, arma::vec &weights)
Run the clusterer, and then turn the cluster assignments into Gaussians.
InitialClusteringType clusterer
Object which will perform the clustering.
Definition: em_fit.hpp:165
double tolerance
Tolerance for convergence of EM.
Definition: em_fit.hpp:163
void Estimate(const arma::mat &observations, std::vector< arma::vec > &means, std::vector< arma::mat > &covariances, arma::vec &weights)
Fit the observations to a Gaussian mixture model (GMM) using the EM algorithm.