-
Notifications
You must be signed in to change notification settings - Fork 0
/
final.m
196 lines (137 loc) · 4.77 KB
/
final.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
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
%% my course_review data
data_raw = reviews(:,[1 2 4])';
max_value = 5;
%% normalize my course_review data
data_raw = normalize_reviews(reviews');
data_raw = data_raw([1 2 4], :);
max_value = 5;
%% make a super dense dataset!
limit = 4000;
user_limit = limit;
item_limit = limit / 2;
R = dense_R(data_raw, item_limit, user_limit, 100, 5);
display(sprintf('%d items, %d users. avg %f reviews per item, %f per user', size(R,1), size(R,2), mean(sum(R > 0,2)), mean(sum(R > 0,1))));
%% INIT DATA
% generate sparse matrix from input data
% just pick the first N points
% NOTE: convert all to double so sparse matrix doesn't complain
% data = double(int0(:,1:20000));
limit = 4000;
user_limit = limit;
item_limit = limit / 2;
under_limit = (data_raw(2,:) < item_limit) & (data_raw(1,:) < user_limit);
data = data_raw(:,under_limit);
num_results = size(data,2);
collision_ratio = num_results / limit;
display(sprintf('found %d results w/ limit %d. collision ratio: %f', num_results, limit, collision_ratio));
% num_nonzeros = sum(data(3,:) > 0);
nonzero_rows = data(3,:) ~= 0;
length(find(nonzero_rows));
data = data(:,nonzero_rows);
% data format: userid, bookid, rating
% want: Rs = IxU matrix
% num_users = length(unique(data(1,:)));
% num_books = length(unique(data(2,:)));
num_users = max(data(1,:));
num_books = max(data(2,:));
% note: max(user) and max(books) are both smaller than 10k
collapse_factor = 1;
I = ceil(num_books / collapse_factor + 1);
U = num_users + 1;
R = zeros(I, U);
for j = 1:size(data,2)
col = data(:,j);
row_i = floor(col(2) / collapse_factor + 1); % item
col_i = col(1) + 1; % user
% note: matlab doesn't like indexing from 0
R(row_i, col_i) = col(3);
end
display(sprintf('avg reviews per item: %f, avg reviews per user: %f', mean(sum(R > 0,2)), mean(sum(R > 0,1))));
clear 'data_raw'; % remove data from memory. maybe this is a bad idea?
clear 'data';
clear 'nonzero_rows';
%% PREP THE LOOP
k = 10; % for k-fold cross-validation
svd_err = zeros(k,2); % row 1: rms, row 2: mae
base_err = zeros(k,2);
% permute the indices before partitioning for cross-validation
nonzero_indices = find(R(:));
i_count = length(nonzero_indices);
indices = nonzero_indices(randperm(i_count));
f = 20; % number of latent dimensions
i = 1;
file_counter = 0;
%% RUN THE LOOP
% gamma = [0.007, 0.01, 0.014] % large gamma => divergence (papers agree)
% lambda = [0.04, 0.1, 0.2, 0.3]
max_iter = 50;
noise = 0.4;
% params: lambda & gamma
% gammas = [.004 .008 .01];
gammas = [.008 .01 .02];
gammas = [0.2 0.3]
lambdas = [.00001 .0001 .001 .01 .1 .2];
errors = zeros(numel(gammas), numel(lambdas)); % IGNORES k right now
base_errors = zeros(1, numel(lambdas));
%%
display(sprintf('starting loop. k = %d, f = %d, max_iter = %d',k, f, max_iter));
% for gamma = [0.001, 0.005, 0.007, 0.01, 0.02, 0.05];
for i = 1:1
%% prep the loop
train = R;
low = 1 + floor((i_count * (i-1))/k);
hi = min(i_count, floor((i_count * i)/k));
test_indices = indices(low:hi);
train(test_indices) = 0; % zero the test indices
R_base = baseline(train, 2);
%%
rmse_b = rms_error(R_base, R, test_indices);
%%
maee_b = mae_error(R_base, R, test_indices);
base_errors(i) = rmse_b;
display(sprintf(' -- i = %d, BASE { rms: %f, mae: %f }', i, rmse_b, maee_b));
%%
for G = 1:numel(gammas)
gamma = gammas(G);
%%
display(sprintf('gamma %f', gamma));
for L = 1:numel(lambdas)
%% run & evaluate
lambda = lambdas(L);
lambda_b = lambda * 2;
%%
R_svd = rsvd_incremental(f, lambda, lambda_b, gamma, max_iter, train, 1, max_value, noise);
rmse = rms_error(R_svd, R, test_indices);
maee = mae_error(R_svd, R, test_indices);
errors(G, L) = rmse;
%%
display(sprintf('lambda %f. improvements: ( %f , %f ) -- SVD {rms: %f, mae: %f}',lambda,rmse_b - rmse, maee_b - maee, rmse,maee));
end
end
end
%%
file_counter = file_counter + 1;
spec_filename = sprintf('./run_data_%d.mat', file_counter );
filename = './run_data.mat';
% run 'load(filename);' to reload lost data
%%
berrors = base_errors;
rerrors = errors;
% backup to disk
save(filename, 'lambdas', 'gammas', 'berrors', 'rerrors');
save(spec_filename, 'lambdas', 'gammas', 'berrors', 'rerrors');
%%
baseline = berrors(1);
%% average errors along the k instances
avg_errors = mean(rerrors,1);
baseline = mean(berrors) * ones(size(lambdas));
plot(lambdas, [avg_errors; baseline]);
%% plot the results (raw!)
err_offset = rerrors - baseline;
plot(lambdas, rerrors);
%%
% calculate and plot avg offset (bigger is better)
e_offset = repmat(berrors', 1, size(rerrors,2)) - rerrors;
plot(lambdas, [e_offset; mean(e_offset,1)]);
xlabel('lambda (regularization term)');
ylabel('RMSE improvement');