forked from ICB-DCM/PESTO
-
Notifications
You must be signed in to change notification settings - Fork 0
/
getPropertySamples.m
176 lines (154 loc) · 5.32 KB
/
getPropertySamples.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
function [properties,fh] = getPropertySamples(properties, parameters, varargin)
% getPropertySamples.m evaluates the properties for the sampled parameters.
%
% USAGE:
% [...] = getPropertySamples(properties,parameters)
% [...] = getPropertySamples(properties,parameters,options)
% [parameters,fh] = getPropertySamples(...)
%
% getPropertySamples() uses the following PestoOptions members:
% * PestoOptions::property_index
% * PestoOptions::mode
% * PestoOptions::fh
% * PestoOptions::save
% * PestoOptions::foldername
% * PestoOptions::comp_type
% * PestoOptions::plot_options
% * PestoOptions::MCMC.thinning
%
% Parameters:
% properties: property struct
% parameters: parameter struct
% varargin:
% options: A PestoOptions object holding various options for the
% algorithm.
%
% Required fields of properties:
% number: number of parameter
% min: lower bound for property values
% max: upper bound for property values
% name: = {'name1',...} ... names of the parameters
% function: = {'function1',...} ... functions to evaluate property
% values. These functions provide the values of the respective
% properties and the corresponding 1st and 2nd order
% derivatives.
%
% Required fields of parameters:
% S: parameter and posterior sample.
% logPost ... log-posterior function along chain
% par ... parameters along chain
% *Note* This struct is obtained using getSamples.m.
%
% Return values:
% properties: updated parameter object
% fh: figure handle
%
% Generated fields of properties:
% S: properties for sampling results
% * par(*,i): ith samples parameter vector
% * logPost(i): log-posterior for ith samples parameter vector
% * prop(j,i): values of jth property for ith samples parameter vector
% * prop_Sigma(*,*,i): covariance of properties for ith samples
% parameter vector
%
% History:
% * 2015/04/01 Jan Hasenauer
% * 2016/10/04 Daniel Weindl
%% Check and assign inputs
if length(varargin) >= 1
options = handleOptionArgument(varargin{1});
else
options = PestoOptions();
end
properties = propertySanityCheck(properties);
% Check initial guess
if ~isfield(parameters,'guess')
parameters.guess = [];
end
% Check and assign options
options.property_index = 1:properties.number;
%% Initialization and figure generation
fh = [];
switch options.mode
case 'visual'
if (isempty(options.fh) || ~isvalid(options.fh))
fh = figure('Name','getPropertySamples');
else
fh = figure(options.fh);
end
case 'text'
fprintf(' \nProperty evaluation:\n====================\n');
end
%% Initialization
properties.S.par = parameters.S.par;
properties.S.logPost = parameters.S.logPost;
properties.S.prop = nan(properties.number,length(properties.S.logPost));
%% Preperation of folder
if options.save
rmdir(options.foldername,'s');
mkdir(options.foldername);
save([options.foldername '/properties_init'],'properties');
end
%% Evaluation of properties for multi-start results -- SEQUENTIAL
if strcmp(options.comp_type,'sequential')
% Loop: Multi-start results
for j = 1:length(properties.S.logPost)
% Loop: Properties
for i = options.property_index
properties.S.prop(i,j) = properties.function{i}(properties.S.par(:,j));
end
% Save
if options.save
dlmwrite([options.foldername '/properties_S' num2str(i,'%d') '__prop.csv'],properties.S.prop(:,j),'delimiter',',','precision',12);
end
% Output
if (mod(j,ceil(length(properties.S.logPost)/100)) == 0) || (j == length(properties.S.logPost))
str = ['Property evaluation for MCMC sampling completed to ' num2str(100*j/length(properties.S.logPost),'%d') ' %'];
switch options.mode
case 'visual', fh = plotPropertySamples(properties,'1D',fh,options.property_index,options.plot_options);
case 'text', disp(str);
case 'silent' % no output
end
end
end
% Output
switch options.mode
% Set the correct options
case 'visual'
options.plot_options.S.plot_type = 1;
fh = plotPropertySamples(properties,'1D',fh,options.property_index,options.plot_options);
end
end
%% Evaluation of properties for multi-start results -- PARALLEL
if strcmp(options.comp_type, 'parallel')
% Initialization
prop = nan(properties.number,length(properties.S.logPost));
% Create local partial copies of the propertry struct
prop_num = properties.number;
prop_fun = properties.function;
prop_S_par = properties.S.par;
opt_save = options.save;
opt_ind = options.property_index;
opt_folder = options.foldername;
% Loop: Multi-start results
parfor i = 1:length(properties.S.logPost)
% Loop: Properties
P = nan(prop_num, 1);
for j = opt_ind
P(j) = prop_fun{j}(prop_S_par(:,i));
end
prop(:,i) = P;
% Save
if (opt_save)
dlmwrite([opt_folder '/properties_S' num2str(i,'%d') '__prop.csv'],prop(:,i),'delimiter',',','precision',12);
end
end
% Assignment
properties.S.prop = prop;
end
%% Output
switch options.mode
case {'visual','text'}, disp('-> Property evaluation for samples FINISHED.');
case 'silent' % no output
end
end