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

Correction on a bug and extension of the original paper for the calculation of incremental weights.

rm-particles^2
Frédéric Karamé 2015-10-07 15:23:42 +02:00
parent 5b5e88b50d
commit 18059c01a6
2 changed files with 13 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/conditional_filter_proposal.m
View File

@ -95,6 +95,7 @@ if ParticleOptions.proposal_approximation.cubature || ParticleOptions.proposal_a
PredictedObservedMean = sum(PredictedObservedMean,2) ;
dState = bsxfun(@minus,tmp(mf0,:),PredictedStateMean)'.*sqrt(weights) ;
dObserved = bsxfun(@minus,tmp(mf1,:),PredictedObservedMean)'.*sqrt(weights);
PredictedStateVariance = dState*dState';
big_mat = [dObserved dState; [H_lower_triangular_cholesky zeros(number_of_observed_variables,number_of_state_variables)] ];
[mat1,mat] = qr2(big_mat,0);
mat = mat';
@ -112,12 +113,13 @@ else
KalmanFilterGain = PredictedStateAndObservedCovariance/PredictedObservedVariance ;
StateVectorMean = PredictedStateMean + KalmanFilterGain*(obs - PredictedObservedMean);
StateVectorVariance = PredictedStateVariance - KalmanFilterGain*PredictedObservedVariance*KalmanFilterGain';
StateVectorVariance = .5*(StateVectorVariance+StateVectorVariance');
StateVectorVarianceSquareRoot = chol(StateVectorVariance)';%reduced_rank_cholesky(StateVectorVariance)';
%StateVectorVariance = .5*(StateVectorVariance+StateVectorVariance');
StateVectorVarianceSquareRoot = chol(StateVectorVariance + 1e-6)' ;
end
PredictedStateVarianceSquareRoot = chol(PredictedStateVariance + 1e-6)' ;
ProposalStateVector = StateVectorVarianceSquareRoot*randn(size(StateVectorVarianceSquareRoot,2),1)+StateVectorMean ;
ypred = measurement_equations(ProposalStateVector,ReducedForm,ThreadsOptions) ;
foo = H_lower_triangular_cholesky \ (obs - ypred) ;
likelihood = exp(-0.5*(foo)'*foo)/normconst2 + 1e-99 ;
Weights = SampleWeights.*likelihood;
Prior = probability2(PredictedStateMean,PredictedStateVarianceSquareRoot,ProposalStateVector) ;
Posterior = probability2(StateVectorMean,StateVectorVarianceSquareRoot,ProposalStateVector) ;
Likelihood = probability2(obs,H_lower_triangular_cholesky,measurement_equations(ProposalStateVector,ReducedForm,ThreadsOptions)) ;
Weights = SampleWeights.*Likelihood.*(Prior./Posterior) ;

16
src/conditional_particle_filter.m
View File

@ -68,10 +68,8 @@ end
% Set persistent variables.
if isempty(init_flag)
%mf0 = ReducedForm.mf0;
mf1 = ReducedForm.mf1;
sample_size = size(Y,2);
%number_of_state_variables = length(mf0);
number_of_observed_variables = length(mf1);
init_flag = 1;
number_of_particles = ParticleOptions.number_of_particles ;
@ -84,25 +82,23 @@ if isempty(H)
H = 0;
H_lower_triangular_cholesky = 0;
else
H_lower_triangular_cholesky = chol(H)'; %reduced_rank_cholesky(H)';
H_lower_triangular_cholesky = chol(H)';
end
% Get initial condition for the state vector.
StateVectorMean = ReducedForm.StateVectorMean;
StateVectorVarianceSquareRoot = reduced_rank_cholesky(ReducedForm.StateVectorVariance)';
StateVectorVarianceSquareRoot = chol(ReducedForm.StateVectorVariance)';
state_variance_rank = size(StateVectorVarianceSquareRoot,2);
Q_lower_triangular_cholesky = chol(Q)'; %reduced_rank_cholesky(Q)';
Q_lower_triangular_cholesky = chol(Q)';
% Set seed for randn().
set_dynare_seed('default');
% Initialization of the likelihood.
normconst2 = log(2*pi)*number_of_observed_variables*prod(diag(H_lower_triangular_cholesky)) ;
normconst2 = sqrt( ((2*pi)^number_of_observed_variables)*prod(det(H)) ) ;
lik = NaN(sample_size,1);
LIK = NaN;
ks = 0 ;
StateParticles = bsxfun(@plus,StateVectorVarianceSquareRoot*randn(state_variance_rank,number_of_particles),StateVectorMean);
SampleWeights = ones(1,number_of_particles)/number_of_particles ;
for t=1:sample_size
@ -120,6 +116,4 @@ for t=1:sample_size
end
end
LIK = -sum(lik(start:end));
LIK = -sum(lik(start:end));