-
Notifications
You must be signed in to change notification settings - Fork 0
/
algos.cpp
410 lines (318 loc) · 9.74 KB
/
algos.cpp
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
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
#include <limits>
#include <string>
#include <fstream>
#include <iomanip>
#include <iostream>
#include <stdlib.h>
#include <armadillo>
#include <random>
using namespace std;
using namespace arma;
#define DBL_MAX numeric_limits<double>::max()
void binovnmtf(const mat& X, const int& k, const int& l, const int& num_iter) {
int n = X.n_rows;
int m = X.n_cols;
mat U(n, k);
mat S(k, l);
cube V(m, l, k);
std::random_device rd;
std::mt19937 gen(rd());
std::uniform_real_distribution<double> unif(0.0, 1.0);
std::uniform_int_distribution<int> unifint(0, 1);
std::uniform_int_distribution<int> unifinttom(0, l-1);
double* U_raw = U.memptr();
double* S_raw = S.memptr();
double* V_raw = V.memptr();
for(int i=0; i < n*k; i++) U_raw[i] = unifint(gen);
for(int i=0; i < k*l; i++) S_raw[i] = unif(gen);
for(int i=0; i < m*l*k; i++) V_raw[i] = unifint(gen);
mat U_best;
U_best.zeros(n, k);
mat S_best;
S_best.zeros(k, l);
cube V_best;
V_best.zeros(m, l, k);
std::cout << "U S V created" << std::endl;
mat X_tilde;
X_tilde.zeros(n, m);
// mat U_tilde;
// U_tilde.zeros(n, l);
// mat V_new;
// V_new.zeros(m, l);
mat V_tilde;
V_tilde.zeros(k, m);
mat V_tilde_best;
V_tilde_best.zeros(k, m);
// mat U_new;
// U_new.zeros(m, k);
rowvec errors_v;
errors_v.zeros(l);
rowvec errors_u;
zeros(k);
double error_best = DBL_MAX;
double error_ant = DBL_MAX;
double error = DBL_MAX;
for (int iter_idx = 0; iter_idx < num_iter; iter_idx++) {
// Compute S
std::cout << "Computing S...";
for (int i=0; i < k; i++) {
for (int j=0; j < l; j++) {
uvec observations_row_cluster_i = find(U.col(i) == 1.0);
uvec observations_col_cluster_j = find(V.slice(i).col(j) == 1.0);
if (observations_row_cluster_i.is_empty() || observations_col_cluster_j.is_empty())
S(i, j) = 0.0;
else
S(i, j) = mean(mean(X.submat( observations_row_cluster_i, observations_col_cluster_j )));
}
}
std::cout << " done!" << std::endl;
// Compute V
std::cout << "Computing V..." << std::endl;
for (int i=0; i < k; i++) {
std::cout << " - for cluster " << i << "...";
// uvec observations_cluster_i = find(U.col(i) == 1.0);
// if (observations_cluster_i.is_empty()) {
// cout << "observations_cluster_i empty!!!!" << endl;
// for (int j=0; j < m; j++)
// V.slice(i)(j, unifinttom(gen)) = 1;
// continue;
// }
mat U_tilde = U.col(i) * S.row(i);
// mat X_tilde = X.rows( observations_cluster_i );
mat X_tilde = X;
V.slice(i).zeros(m, l);
for (int j=0; j < m; j++) {
errors_v.zeros(l);
for (int col_clust_idx = 0; col_clust_idx < l; col_clust_idx++)
errors_v(col_clust_idx) = sum(pow(X_tilde.col(j) - U_tilde.col(col_clust_idx), 2));
uword ind;
errors_v.min(ind);
V.slice(i)(j, ind) = 1;
}
std::cout << " done!" << std::endl;
}
std::cout << "Done computed V!" << std::endl;
std::cout << "Computing U...";
// Compute V_tilde
V_tilde.zeros(k, m);
for (int i=0; i < k; i++)
V_tilde.row(i) = S.row(i) * V.slice(i).t();
U.zeros(n, k);
for (int i = 0; i < n; i++) {
errors_u.zeros(k);
for (int row_clust_idx = 0; row_clust_idx < k; row_clust_idx++)
errors_u(row_clust_idx) =
sum(pow(X.row(i) - V_tilde.row(row_clust_idx), 2));
uword ind;
errors_u.min(ind);
U(i, ind) = 1;
}
std::cout << " done!" << std::endl;
error_ant = error;
error = accu(pow(X - (U * V_tilde), 2));
std::cout << "Error obtained: " << error << std::endl;
if (error < error_best) {
U_best = U;
S_best = S;
V_best = V;
V_tilde_best = V_tilde;
error_best = error;
}
double precision_err = 0.0001;
if (abs(error - error_ant) <= precision_err) break;
}
mat Reconstruction = U_best * V_tilde_best;
U_best.save("U.csv", csv_ascii);
S_best.save("S.csv", csv_ascii);
V_best.save("V.hdf5", hdf5_binary);
Reconstruction.save("Reconstruction.csv", csv_ascii);
ofstream errorfile("error.csv");
errorfile << std::fixed << std::setprecision(8) << error_best;
errorfile.close();
}
void fnmtf(const mat& X, const int& k, const int& l, const int& num_iter) {
int m = X.n_rows;
int n = X.n_cols;
mat U(m, k);
mat S(k, l);
mat V(n, l);
std::random_device rd;
std::mt19937 gen(rd());
std::uniform_real_distribution<double> unif(0.0, 1.0);
double* U_raw = U.memptr();
double* S_raw = S.memptr();
double* V_raw = V.memptr();
for(int i=0; i < m*k; i++) U_raw[i] = unif(gen);
for(int i=0; i < k*l; i++) S_raw[i] = unif(gen);
for(int i=0; i < n*l; i++) V_raw[i] = unif(gen);
// U.print("U before: ");
// S.print("S before: ");
// V.print("V before: ");
mat U_best;
U_best.zeros(m, k);
mat S_best;
S_best.zeros(k, l);
mat V_best;
V_best.zeros(n, l);
std::cout << "U S V created" << std::endl;
// U.print("U:");
// S.print("S:");
// V.print("V:");
mat U_tilde;
U_tilde.zeros(m, l);
mat V_new;
V_new.zeros(n, l);
mat V_tilde;
V_tilde.zeros(k, m);
mat U_new;
U_new.zeros(m, k);
rowvec errors_v;
errors_v.zeros(l);
rowvec errors_u;
double error_best = DBL_MAX;
double error_ant = DBL_MAX;
double error = DBL_MAX;
for (int iter_idx = 0; iter_idx < num_iter; iter_idx++) {
std::cout << "Computing S...";
S = pinv(U.t() * U) * (U.t() * (X * V)) * pinv(V.t() * V);
std::cout << " OK!\n";
std::cout << "Computing V...";
U_tilde = U * S;
V_new.zeros(n, l);
for (int j = 0; j < n; j++) {
errors_v.zeros(l);
for (int col_clust_idx = 0; col_clust_idx < l; col_clust_idx++)
errors_v(col_clust_idx) =
sum(pow(X.col(j) - U_tilde.col(col_clust_idx), 2));
uword ind;
errors_v.min(ind);
V_new(j, ind) = 1;
}
V = V_new;
std::cout << " OK!\n";
std::cout << "Computing U...";
V_tilde = S * V.t();
U_new.zeros(m, k);
for (int i = 0; i < m; i++) {
errors_u.zeros(k);
for (int row_clust_idx = 0; row_clust_idx < k; row_clust_idx++)
errors_u(row_clust_idx) =
sum(pow(X.row(i) - V_tilde.row(row_clust_idx), 2));
uword ind;
errors_u.min(ind);
U_new(i, ind) = 1;
}
U = U_new;
std::cout << " OK!\n";
error_ant = error;
error = accu(pow(X - (U * S * V.t()), 2));
std::cout << "Error obtained: " << error << "\n";
if (error < error_best) {
U_best = U;
S_best = S;
V_best = V;
error_best = error;
}
double precision_err = 0.000001;
if (abs(error - error_ant) <= precision_err) break;
}
U_best.save("U.csv", csv_ascii);
S_best.save("S.csv", csv_ascii);
V_best.save("V.csv", csv_ascii);
ofstream errorfile("error.csv");
errorfile << std::fixed << std::setprecision(8) << error_best;
errorfile.close();
}
void onmtf(const mat& X, const int& k, const int& l, const int& num_iter) {
int m = X.n_rows;
int n = X.n_cols;
mat U(m, k);
mat S(k, l);
mat V(n, l);
std::random_device rd;
std::mt19937 gen(rd());
std::uniform_real_distribution<double> unif(0.0, 1.0);
double* U_raw = U.memptr();
double* S_raw = S.memptr();
double* V_raw = V.memptr();
for(int i=0; i < m*k; i++) U_raw[i] = unif(gen);
for(int i=0; i < k*l; i++) S_raw[i] = unif(gen);
for(int i=0; i < n*l; i++) V_raw[i] = unif(gen);
// U.print("U before: ");
// S.print("S before: ");
// V.print("V before: ");
mat U_best;
U_best.zeros(m, k);
mat S_best;
S_best.zeros(k, l);
mat V_best;
V_best.zeros(n, l);
mat U_norm;
U_norm.zeros(m, k);
mat V_norm;
V_norm.zeros(n, l);
std::cout << "U S V created" << std::endl;
double error_best = DBL_MAX;
double error_ant = DBL_MAX;
double error = DBL_MAX;
for (int iter_idx = 0; iter_idx < num_iter; iter_idx++) {
U = U % ((X * V * S.t()) / (U * S * V.t() * X.t() * U));
V = V % ((X.t() * U * S) / (V * S.t() * U.t() * X * V));
S = S % ((U.t() * X * V) / (U.t() * U * S * V.t() * V));
error_ant = error;
error = accu(pow((X - U * S * V.t()), 2));
std::cout << "Error obtained: " << error << "\n";
if (error < error_best) {
U_best = U;
S_best = S;
V_best = V;
error_best = error;
}
double precision_err = 0.000001;
if (abs(error - error_ant) <= precision_err) break;
}
mat Du = diagmat((ones<rowvec>(m) * U_best));
mat Dv = diagmat((ones<rowvec>(n) * V_best));
// Du.print("Du: ");
// Dv.print("Dv: ");
U_norm = U_best * (diagmat(S_best * Dv * ones<colvec>(l)));
V_norm = V_best * (diagmat(ones<rowvec>(k) * Du * S_best));
U_norm.save("U.csv", csv_ascii);
S_best.save("S.csv", csv_ascii);
V_norm.save("V.csv", csv_ascii);
ofstream errorfile("error.csv");
errorfile << std::fixed << std::setprecision(8) << error_best;
errorfile.close();
}
int main(int argc, char* argv[]) {
if (argc != 5) {
cout << "Missing arguments...\n";
cout << argc;
return 1;
}
cout.precision(11);
cout.setf(ios::fixed);
string algo_name = argv[1];
const int k = atoi(argv[2]);
const int l = atoi(argv[3]);
const int num_iter = atoi(argv[4]);
std::cout << "Reading X...";
mat X;
// X.load("data.csv");
X.load("data.h5", hdf5_binary);
std::cout << " read " << X.n_rows << " rows and " << X.n_cols << " cols!\n";
// X.raw_print(cout, "X: ");
clock_t begin = clock();
if (algo_name == "fnmtf")
fnmtf(X, k, l, num_iter);
else if (algo_name == "onmtf")
onmtf(X, k, l, num_iter);
else if (algo_name == "bin_ovnmtf")
binovnmtf(X, k, l, num_iter);
else
cout << "Wrong algo name...\n";
clock_t end = clock();
double elapsed_secs = double(end - begin) / CLOCKS_PER_SEC;
std::cout << "Time taken: " << elapsed_secs << std::endl;
return 0;
}