Public Member Functions
void	setTransitionProbabilities (const MatrixXd &t)
	Transition matrix is a States x States matrix, where rows are the target state and columns the source states. Defines probability to change from one state to another.

void	setTransitionProbabilitiesCallback (std::function< double(int stateFrom, int stateTo, int timestepFrom, int timestepTo)> callback)
	Set callback which is evaluating the transition probability given the current observation assignments.

void	setObservationProbabilities (const MatrixXd &o)
	Observation matrix is a Observation x States matrix. Defines probability to see an observation from a particular state.

void	setInitialProbabilities (const VectorXd &init)
	Initial probabilities set the probabilities of the state at time t=0.

MatrixXd	estimate (const MatrixXd &observations, MatrixXd *optStateProbs=nullptr)
	Observations should be a matrix of Observations x Time (row/col respectively)

std::vector< int >	viterbi (const MatrixXd &observations, MatrixXd optStateProbs=nullptr, double logProbability=nullptr, Eigen::MatrixXd *transitionMatrix=nullptr, bool normalize=true, bool logViterbi=false)
	Find most likely sequence of states given observations.

std::vector< int >	logViterbi (double *logProbability=nullptr)
	Reduced faster version of the `viterbi` mathod. Same as `viterbi` with square identity `observation` matrix and `logViterbi` set to true.

Member Function Documentation

◆ setTransitionProbabilitiesCallback()

void setTransitionProbabilitiesCallback ( std::function< double(int stateFrom, int stateTo, int timestepFrom, int timestepTo)> callback )

Set callback which is evaluating the transition probability given the current observation assignments.

Parameters

callback Callback When the callback is set the transition probabilities matrix if also set is ignored

The documentation for this class was generated from the following file: