stepan
/
dynare
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

include parallelisation, correction of the likelihood calculation for the Amisano and Tristani option and the possibility of monte carlo draws for the proposal.

rm-particles^2
Frédéric Karamé 2018-06-28 09:21:50 +02:00
parent 35c3a9fcd9
commit fb7ec278d6
1 changed files with 9 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/conditional_filter_proposal.m
View File

@ -74,7 +74,11 @@ if isempty(init_flag2)
init_flag2 = 1;
end
if ParticleOptions.proposal_approximation.cubature || ParticleOptions.proposal_approximation.montecarlo
if ParticleOptions.proposal_approximation.montecarlo
nodes = randn(ParticleOptions.number_of_particles/10,number_of_structural_innovations) ;
weights = 1/ParticleOptions.number_of_particles ;
weights_c = weights ;
elseif ParticleOptions.proposal_approximation.cubature
[nodes,weights] = spherical_radial_sigma_points(number_of_structural_innovations);
weights_c = weights ;
elseif ParticleOptions.proposal_approximation.unscented
@ -108,7 +112,8 @@ if ParticleOptions.proposal_approximation.cubature || ParticleOptions.proposal_a
StateVectorMean = PredictedStateMean + (CovarianceObservedStateSquareRoot/PredictedObservedVarianceSquareRoot)*Error ;
if ParticleOptions.cpf_weights_method.amisanotristani
Weights = SampleWeights.*probability2(zeros(number_of_observed_variables,1),PredictedObservedVarianceSquareRoot,Error) ;
end
% Weights = SampleWeights.*sum(weights*probability2(obs,H_lower_triangular_cholesky,tmp(mf1,:)),1) ;
end
else
dState = bsxfun(@minus,tmp(mf0,:),PredictedStateMean);
dObserved = bsxfun(@minus,tmp(mf1,:),PredictedObservedMean);
@ -130,7 +135,8 @@ else
end
if ParticleOptions.cpf_weights_method.amisanotristani
Weights = SampleWeights.*probability2(zeros(number_of_observed_variables,1),chol(PredictedObservedVariance)',Error) ;
end
% Weights = SampleWeights.*sum(probability2(obs,H_lower_triangular_cholesky,tmp(mf1,:))*weights) ;
end
end
ProposalStateVector = StateVectorVarianceSquareRoot*randn(size(StateVectorVarianceSquareRoot,2),1)+StateVectorMean ;