Inheritance diagram for GenericMetropolis:

Public Member Functions
	GenericMetropolis (const PsiPsychometric Model, const PsiData Data, PsiRandom *proposal)
void	proposePoint (std::vector< double > &current_theta, std::vector< double > &step_widths, PsiRandom *proposal, std::vector< double > &new_theta)
	propose a new sample and save it in new_theta
void	findOptimalStepwidth (PsiMClist const &pilot)
	Find the optimal stepwidth by regressing each parameter against the others.

Constructor & Destructor Documentation

GenericMetropolis::GenericMetropolis	(	const PsiPsychometric *	Model,
		const PsiData *	Data,
		PsiRandom *	proposal
	)			`[inline]`

Parameters:

	Model	psychometric funciton model to sample from
	Data	data to base inference on
	proposal	proposal distribution (will usually be a gaussian)

Member Function Documentation

void GenericMetropolis::findOptimalStepwidth ( PsiMClist const & pilot )

Find the optimal stepwidth by regressing each parameter against the others.

For each parameter, do a regression using QR-decomposition and take the residuals to calculate the optimal stepwidth.

Parameters:

pilot

a pilot sample to base the regression on

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