auxiliary_particle_filter.m 6.38 KB
Newer Older
1
function [LIK,lik] = auxiliary_particle_filter(ReducedForm,Y,start,DynareOptions)
Stéphane Adjemian's avatar
Stéphane Adjemian committed
2

3
4
% Evaluates the likelihood of a nonlinear model with a particle filter allowing eventually resampling.

Stéphane Adjemian's avatar
Stéphane Adjemian committed
5
% Copyright (C) 2011-2014 Dynare Team
6
%
Stéphane Adjemian's avatar
Stéphane Adjemian committed
7
% This file is part of Dynare (particles module).
8
9
10
11
12
13
%
% Dynare is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
Stéphane Adjemian's avatar
Stéphane Adjemian committed
14
% Dynare particles module is distributed in the hope that it will be useful,
15
16
17
18
19
20
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License
% along with Dynare.  If not, see <http://www.gnu.org/licenses/>.
Stéphane Adjemian's avatar
Stéphane Adjemian committed
21

22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
persistent init_flag mf0 mf1 number_of_particles
persistent sample_size number_of_state_variables number_of_observed_variables number_of_structural_innovations

% Set default
if isempty(start)
    start = 1;
end

% Set flag for prunning
pruning = DynareOptions.particle.pruning;

% Get steady state and mean.
steadystate = ReducedForm.steadystate;
constant = ReducedForm.constant;
state_variables_steady_state = ReducedForm.state_variables_steady_state;

% 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);
    number_of_structural_innovations = length(ReducedForm.Q);
    number_of_particles = DynareOptions.particle.number_of_particles;
    init_flag = 1;
end

% Set local state space model (first order approximation).
ghx  = ReducedForm.ghx;
ghu  = ReducedForm.ghu;
% Set local state space model (second order approximation).
ghxx = ReducedForm.ghxx;
ghuu = ReducedForm.ghuu;
ghxu = ReducedForm.ghxu;

% Get covariance matrices
Q = ReducedForm.Q;
H = ReducedForm.H;
if isempty(H)
    H = 0;
end

% Get initial condition for the state vector.
StateVectorMean = ReducedForm.StateVectorMean;
StateVectorVarianceSquareRoot = reduced_rank_cholesky(ReducedForm.StateVectorVariance)';
state_variance_rank = size(StateVectorVarianceSquareRoot,2);
Q_lower_triangular_cholesky = chol(Q)';
if pruning
    StateVectorMean_ = StateVectorMean;
    StateVectorVarianceSquareRoot_ = StateVectorVarianceSquareRoot;
end

% Set seed for randn().
set_dynare_seed('default');

% Initialization of the likelihood.
const_lik = log(2*pi)*number_of_observed_variables;
lik  = NaN(sample_size,1);
LIK  = NaN;

% Initialization of the weights across particles.
weights = ones(1,number_of_particles)/number_of_particles ;
StateVectors = bsxfun(@plus,StateVectorVarianceSquareRoot*randn(state_variance_rank,number_of_particles),StateVectorMean);
if pruning
    StateVectors_ = StateVectors;
end
for t=1:sample_size
    yhat = bsxfun(@minus,StateVectors,state_variables_steady_state);
    if pruning
        yhat_ = bsxfun(@minus,StateVectors_,state_variables_steady_state);
        [tmp, tmp_] = local_state_space_iteration_2(yhat,zeros(number_of_structural_innovations,number_of_particles),ghx,ghu,constant,ghxx,ghuu,ghxu,yhat_,steadystate,DynareOptions.threads.local_state_space_iteration_2);
    else
        tmp = local_state_space_iteration_2(yhat,zeros(number_of_structural_innovations,number_of_particles),ghx,ghu,constant,ghxx,ghuu,ghxu,DynareOptions.threads.local_state_space_iteration_2);
    end
    PredictedObservedMean = weights*(tmp(mf1,:)');
    PredictionError = bsxfun(@minus,Y(:,t),tmp(mf1,:));
    dPredictedObservedMean = bsxfun(@minus,tmp(mf1,:),PredictedObservedMean');
    PredictedObservedVariance = bsxfun(@times,weights,dPredictedObservedMean)*dPredictedObservedMean' +H;
    wtilde = exp(-.5*(const_lik+log(det(PredictedObservedVariance))+sum(PredictionError.*(PredictedObservedVariance\PredictionError),1))) ;
    tau_tilde = weights.*wtilde ;
    sum_tau_tilde = sum(tau_tilde) ;
    %var_wtilde = wtilde-sum_tau_tilde ;
    %var_wtilde = var_wtilde'*var_wtilde/(number_of_particles-1) ;
    lik(t) = log(sum_tau_tilde) ; %+ .5*var_wtilde/(number_of_particles*(sum_tau_tilde*sum_tau_tilde)) ;
    tau_tilde = tau_tilde/sum_tau_tilde;
    if pruning
        temp = resample([yhat' yhat_'],tau_tilde',DynareOptions);
        yhat = temp(:,1:number_of_state_variables)' ;
        yhat_ = temp(:,number_of_state_variables+1:2*number_of_state_variables)' ;
    else
        yhat = resample(yhat',tau_tilde',DynareOptions)' ;
    end
    if pruning
        [tmp, tmp_] = local_state_space_iteration_2(yhat,zeros(number_of_structural_innovations,number_of_particles),ghx,ghu,constant,ghxx,ghuu,ghxu,yhat_,steadystate,DynareOptions.threads.local_state_space_iteration_2);
    else
        tmp = local_state_space_iteration_2(yhat,zeros(number_of_structural_innovations,number_of_particles),ghx,ghu,constant,ghxx,ghuu,ghxu,DynareOptions.threads.local_state_space_iteration_2);
    end
    PredictedObservedMean = weights*(tmp(mf1,:)');
    PredictionError = bsxfun(@minus,Y(:,t),tmp(mf1,:));
    dPredictedObservedMean = bsxfun(@minus,tmp(mf1,:),PredictedObservedMean');
    PredictedObservedVariance = bsxfun(@times,weights,dPredictedObservedMean)*dPredictedObservedMean' +H;
    wtilde = exp(-.5*(const_lik+log(det(PredictedObservedVariance))+sum(PredictionError.*(PredictedObservedVariance\PredictionError),1))) ;
    epsilon = Q_lower_triangular_cholesky*randn(number_of_structural_innovations,number_of_particles);
    if pruning
        [tmp, tmp_] = local_state_space_iteration_2(yhat,epsilon,ghx,ghu,constant,ghxx,ghuu,ghxu,yhat_,steadystate,DynareOptions.threads.local_state_space_iteration_2);
        StateVectors_ = tmp_(mf0,:);
    else
        tmp = local_state_space_iteration_2(yhat,epsilon,ghx,ghu,constant,ghxx,ghuu,ghxu,DynareOptions.threads.local_state_space_iteration_2);
    end
    StateVectors = tmp(mf0,:);
    PredictedObservedMean = mean(tmp(mf1,:),2);
    PredictionError = bsxfun(@minus,Y(:,t),tmp(mf1,:));
    dPredictedObservedMean = bsxfun(@minus,tmp(mf1,:),PredictedObservedMean);
    PredictedObservedVariance = (dPredictedObservedMean*dPredictedObservedMean')/number_of_particles + H;
    lnw = exp(-.5*(const_lik+log(det(PredictedObservedVariance))+sum(PredictionError.*(PredictedObservedVariance\PredictionError),1)));
    wtilde = lnw./wtilde;
    weights = wtilde/sum(wtilde);
end

LIK = -sum(lik(start:end));