Implements a multi-variate gaussian distribution. More...
Inheritance diagram for drwnGaussian:
Public Member Functions | |
| drwnGaussian (int n=1) | |
construct an n dimensional zero-mean identity-covariance gaussian | |
| drwnGaussian (const VectorXd &mu, double sigma2) | |
| construct a gaussian with given mean and isotropic covariance | |
| drwnGaussian (const VectorXd &mu, const MatrixXd &sigma2) | |
| construct a gaussian with given mean and covariance | |
| drwnGaussian (const vector< double > &mu, double sigma2) | |
| construct a gaussian with given mean and isotropic covariance | |
| drwnGaussian (const drwnSuffStats &stats) | |
| construct a gaussian from given second-order sufficient statistics | |
| drwnGaussian (const drwnGaussian &model) | |
| copy constructor | |
| void | initialize (int n) |
initialize the gaussian to be n dimensional zero-mean and identity-covariance | |
| void | initialize (const VectorXd &mu, double sigma2) |
initialize the gaussian to have mean mu and isotropic variance sigma2 | |
| void | initialize (const VectorXd &mu, const MatrixXd &sigma2) |
| initialize the gaussian to the given mean and covariance | |
| drwnGaussian * | marginalize (const vector< int > &indx) const |
Generate a gaussian with the variables not in indx marginalized out. The calling function is responsible for deleting the returned object. | |
| drwnGaussian | reduce (const vector< double > &x, const vector< int > &indx) const |
Returns a gaussian conditioned on x. Repeated calls to reduce on the same set of variables should use the conditionOn function. | |
| drwnGaussian | reduce (const map< int, double > &x) const |
Returns a gaussian conditioned on x. | |
| drwnConditionalGaussian | conditionOn (const vector< int > &indx) const |
| Returns a temporary objects for conditioning on a set of variables. | |
| const char * | type () const |
| returns object type as a string (e.g., Foo::type() { return "Foo"; }) | |
| drwnGaussian * | clone () const |
| returns a copy of the class usually implemented as virtual Foo* clone() { return new Foo(*this); } | |
| bool | save (drwnXMLNode &xml) const |
| write object to XML node (see also write) | |
| bool | load (drwnXMLNode &xml) |
| read object from XML node (see also read) | |
| void | evaluate (const MatrixXd &x, VectorXd &p) const |
compute the log-likelihood of each row of X and put the results in p | |
| void | evaluate (const vector< vector< double > > &x, vector< double > &p) const |
compute the log-likelihood of each vector in x and put the results in p | |
| double | evaluateSingle (const VectorXd &x) const |
| compute the log-likelihood of a given vector | |
| double | evaluateSingle (const vector< double > &x) const |
| compute the log-likelihood of a given vector | |
| double | evaluateSingle (double x) const |
| compute the log-likelihood of a scalar (for one dimensional gaussians only) | |
| void | sample (VectorXd &x) const |
| generate a random sample from the gaussian | |
| void | sample (vector< double > &x) const |
| generate a random sample from the gaussian | |
| void | train (const MatrixXd &x, double lambda=0.0) |
Estimate the mean and covariance of the gaussian from a matrix of training examples. Data should be arranged row-wise. The parameter lambda can be used to regularize the covariance matrix with an additive isotropic component. | |
| void | train (const vector< vector< double > > &x, double lambda=0.0) |
| See above. | |
| void | train (const vector< double > &x, double lambda=0.0) |
| See above, but for one dimensional gaussians. | |
| void | train (const drwnSuffStats &stats, double lambda=0.0) |
| See above, but using given second-order sufficient statistics. | |
| unsigned | dimension () const |
| returns the dimensionality of the gaussian | |
| const VectorXd & | mean () const |
| returns the mean of the gaussian | |
| const MatrixXd & | covariance () const |
| returns the covariance matrix for the gaussian | |
| double | logPartitionFunction () const |
| computes the log partition function of the gaussian | |
| double | klDivergence (const drwnGaussian &model) const |
| computes the KL divergence between the gaussian and the given model | |
| double | klDivergence (const drwnSuffStats &stats) const |
| computes the KL divergence between the gaussian and distribution induced by the given sufficient statistics | |
| drwnGaussian & | operator= (const drwnGaussian &model) |
| assignment operator | |
 Public Member Functions inherited from drwnWriteable | |
| bool | write (const char *filename) const |
| write object to file (calls save) | |
| bool | read (const char *filename) |
| read object from file (calls load) | |
| void | dump () const |
| print object's current state to standard output (for debugging) | |
Static Public Attributes | |
| static bool | AUTO_RIDGE = false |
Protected Member Functions | |
| void | guaranteeInvSigma () const |
| void | freeCachedParameters () |
| void | updateCachedParameters () |
Detailed Description
![\[ \log p(x; \mu, \Sigma) = -\frac{n}{2} \log(2 \pi) -\frac{1}{2} \log |\Sigma| - \frac{1}{2}\left(x - \mu\right)^T \Sigma^{-1} \left(x - \mu\right) \]](form_31.png)
The documentation for this class was generated from the following files:
- drwnGaussian.h
- drwnGaussian.cpp
