forked from HIIT/knowledge-elicitation-for-linear-regression
-
Notifications
You must be signed in to change notification settings - Fork 0
/
main_all.m
166 lines (161 loc) · 9.67 KB
/
main_all.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
close all
clear all
RNG_SEED = rng;
%% Parameters and Simulator setup
MODE = 1;
% MODE specifies the type of feedback and the model that we are using
% 1: Feedback on weight values. Model: spike and slab prior
% 2: Feedback on relevance of features. Model: spike and slab prior
%data parameters for simulation data
num_features = 12:2:200; %[start,step,max] This can be a set of values (e.g. 1:100) or just one value (e.g. 100)
num_trainingdata = 10; %[start,step,max] This can be a set of values (e.g. 2:1:100) or just one value (e.g. 5)
num_userdata = 500; %data that will be used in active learning
max_num_nonzero_features = 10; % maximum number of features that are nonzero --- AKA sparsity measure
%Algorithm parameters
num_iterations = 50; %total number of user feedback
num_runs = 100; %total number of runs (necessary for averaging results)
%model parameters
sparse_options = struct('damp',0.8, 'damp_decay',0.95, 'robust_updates',2, 'verbosity',0, 'max_iter',1000, 'threshold',1e-5, 'min_site_prec',1e-6);
sparse_params = struct('sigma2',1^2, 'tau2', 1^2 ,'eta2',0.1^2,'p_u', 0.95, 'simulated_data', 1);
%% METHOD LIST
% Set the desirable methods to 'True' and others to 'False'. only the 'True' methods will be considered in the simulation
METHODS_ED = {
'True', 'Random';
'False', 'First relevant features, then non-relevant';
'False', 'Max posterior inclusion probability';
'False', 'max variance';
'True', 'Expected information gain, full EP approx';
'False', 'Expected information gain, full EP approx, non-sequential';
'False', 'Expected information gain, fast approx'; %fast approx methdos are available for MODE = 2 only
'False', 'Expected information gain, fast approx, non-sequential' %fast approx methdos are available for MODE = 2 only
};
METHODS_AL = {
'False', 'AL:Uniformly random';
'False', 'AL: Expected information gain'
};
METHODS_GT = {
'False', 'Ground truth - all data';
'False', 'Ground truth - all feedback'
};
Method_list_ED = [];
for m = 1:size(METHODS_ED,1)
if strcmp(METHODS_ED(m,1),'True')
Method_list_ED = [Method_list_ED,METHODS_ED(m,2)];
end
end
Method_list_AL = [];
for m = 1:size(METHODS_AL,1)
if strcmp(METHODS_AL(m,1),'True')
Method_list_AL = [Method_list_AL,METHODS_AL(m,2)];
end
end
Method_list_GT = [];
for m = 1:size(METHODS_GT,1)
if strcmp(METHODS_GT(m,1),'True')
Method_list_GT = [Method_list_GT,METHODS_GT(m,2)];
end
end
Method_list = [Method_list_GT, Method_list_ED, Method_list_AL];
num_methods = size(Method_list,2); %number of decision making methods that we want to consider
%% Main
Loss_1 = zeros(num_methods, num_iterations, num_runs, size(num_features,2),size(num_trainingdata,2));
Loss_2 = zeros(num_methods, num_iterations, num_runs, size(num_features,2),size(num_trainingdata,2));
decisions = zeros(num_methods, num_iterations, num_runs, size(num_features,2),size(num_trainingdata,2));
tic
for n_f = 1:size(num_features,2);
disp(['For feature index ', num2str(n_f), ' out of ', num2str(size(num_features,2)), '. acc time = ', num2str(toc) ]);
sparse_params.rho = max_num_nonzero_features/num_features(n_f);
for n_t = 1:size(num_trainingdata,2);
disp(['For training index ', num2str(n_t), 'out of ', num2str(size(num_trainingdata,2)), '. acc time = ', num2str(toc) ]);
num_data = 500 + num_trainingdata(n_t) + num_userdata; % total number of data (training and test)
for run = 1:num_runs
disp(['run number ', num2str(run), ' from ', num2str(num_runs), '. acc time = ', num2str(toc) ]);
num_nonzero_features = min( num_features(n_f), max_num_nonzero_features);
%Theta_star is the true value of the unknown weight vector
% non-zero elements of theta_star are generated based on the model parameters
theta_star = sqrt(sparse_params.tau2)*randn( num_nonzero_features, 1);
theta_star = [theta_star; zeros(num_features(n_f)-num_nonzero_features,1)]; % make it sparse
z_star = theta_star ~= 0; % the true value for the latent variable Z in spike and slab model
%generate new data for each run (because the results is sensitive to the covariate values)
X_all = mvnrnd(zeros(num_features(n_f),1), 1.0*eye(num_features(n_f),num_features(n_f)),num_data);
Y_all = normrnd(X_all*theta_star, sqrt(sparse_params.sigma2));
[X_train, X_user, X_test, Y_train, Y_user, Y_test] = partition_data(X_all, Y_all, num_userdata, num_trainingdata(n_t));
%% main algorithms (ED, AL, and GT)
for method_num = 1:num_methods
method_name = Method_list(method_num);
%Feedback = values (1st column) and indices (2nd column) of user feedback
Feedback = []; %only used in experimental design methdos
%selected_data = indices of data selected by active learning from X_user and Y_user
selected_data = []; %only used in active learning methods
sparse_options.si = []; %carry prior site terms between interactions
%% Calculate ground truth solutions
if find(strcmp(Method_list_GT, method_name))
if find(strcmp('Ground truth - all data', method_name))
%calculate the posterior based on all train+user data
posterior = calculate_posterior([X_train, X_user], [Y_train; Y_user], Feedback, ...
MODE, sparse_params, sparse_options);
end
if find(strcmp('Ground truth - all feedback', method_name))
%calculate the posterior based on all feedbacks
for feature_index = 1:size(X_train,1)
new_fb_value = user_feedback(feature_index, theta_star, z_star, MODE, sparse_params);
Feedback = [Feedback; new_fb_value , feature_index];
end
posterior = calculate_posterior(X_train, Y_train, Feedback, ...
MODE, sparse_params, sparse_options);
end
Y_hat = X_test'*posterior.mean;
Y_hat_train = X_train'*posterior.mean;
Loss_1(method_num, :, run, n_f ,n_t) = mean((Y_hat- Y_test).^2);
Loss_2(method_num, :, run, n_f ,n_t) = mean((Y_hat_train- Y_train).^2);
continue
end
%% for non-sequential ED methods find the suggested queries before user interaction
if strfind(char(method_name),'non-sequential')
posterior = calculate_posterior(X_train, Y_train, [], MODE, sparse_params, sparse_options);
%find non-sequential order of features to be queried from the user
non_seq_feature_indices = decision_policy(posterior, method_name, z_star, X_train, Y_train, ...
[], MODE, sparse_params, sparse_options);
end
%% User interaction
for it = 1:num_iterations %number of user feedback
%calculate the posterior based on training + feedback until now
posterior = calculate_posterior([X_train, X_user(:,selected_data)], [Y_train; Y_user(selected_data)], Feedback, ...
MODE, sparse_params, sparse_options);
sparse_options.si = posterior.si;
%% calculate different loss functions
Y_hat = X_test'*posterior.mean;
Y_hat_train = X_train'*posterior.mean;
Loss_1(method_num, it, run, n_f ,n_t) = mean((Y_hat- Y_test).^2);
Loss_2(method_num, it, run, n_f ,n_t) = mean((Y_hat_train- Y_train).^2);
%% If ED: make a decision based on ED decision policy
if find(strcmp(Method_list_ED, method_name))
%for non-sequential methods, use the saved order
if strfind(char(method_name),'non-sequential')
feature_index = non_seq_feature_indices(it);
else
%for sequential methods find the next decision based on feedback until now
feature_index = decision_policy(posterior, method_name, z_star, X_train, Y_train, ...
Feedback, MODE, sparse_params, sparse_options);
end
decisions(method_num, it, run, n_f ,n_t) = feature_index;
%simulate user feedback
new_fb_value = user_feedback(feature_index, theta_star, z_star, MODE, sparse_params);
Feedback = [Feedback; new_fb_value , feature_index];
end
%% If AL: add a new data point based on AL decision policy
if find(strcmp(Method_list_AL, method_name))
[new_selected_data] = decision_policy_AL(posterior, method_name, ...
[X_train, X_user(:,selected_data)] , [Y_train; Y_user(selected_data)], ...
X_user, selected_data, sparse_params, sparse_options);
selected_data = [selected_data;new_selected_data];
end
end
end
end
end
end
%% averaging and plotting
save('results_all', 'Loss_1', 'Loss_2', 'decisions', 'sparse_options', 'sparse_params', ...
'z_star', 'Method_list', 'num_features','num_trainingdata', 'MODE', 'RNG_SEED')
evaluate_results_all