Datasets:

Modalities:
Tabular
Text
Formats:
csv
ArXiv:
DOI:
Libraries:
Datasets
pandas
License:
pfilonenko commited on
Commit
24d8f62
·
verified ·
1 Parent(s): 4574c24

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +179 -11
README.md CHANGED
@@ -70,6 +70,19 @@ In these files, there are following fields:
70
  - **SKLEARN_LogReg_test** is a statistic of the proposed ML-method based on Logistic Regression (implemented in sklearn);
71
  - **SKLEARN_GB_test** is a statistic of the proposed ML-method based on Gradient Boosting Machine (implemented in sklearn).
72
 
 
 
 
 
 
 
 
 
 
 
 
 
 
73
  # Dataset Simulation
74
 
75
  For this dataset, the full source code (C++) is available [here](https://github.com/pfilonenko/ML_for_TwoSampleTesting/tree/main/dataset/simulation).
@@ -166,16 +179,171 @@ int main()
166
  return 0;
167
  }
168
  ```
 
 
 
 
169
 
170
- # Citing
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
171
 
172
- ~~~
173
- @misc {petr_philonenko_2024,
174
- author = { {Petr Philonenko} },
175
- title = { ML_for_TwoSampleTesting (Revision a4ae672) },
176
- year = 2024,
177
- url = { https://huggingface.co/datasets/pfilonenko/ML_for_TwoSampleTesting },
178
- doi = { 10.57967/hf/2978 },
179
- publisher = { Hugging Face }
180
- }
181
- ~~~
 
70
  - **SKLEARN_LogReg_test** is a statistic of the proposed ML-method based on Logistic Regression (implemented in sklearn);
71
  - **SKLEARN_GB_test** is a statistic of the proposed ML-method based on Gradient Boosting Machine (implemented in sklearn).
72
 
73
+ # Citing
74
+
75
+ ~~~
76
+ @misc {petr_philonenko_2024,
77
+ author = { {Petr Philonenko} },
78
+ title = { ML_for_TwoSampleTesting (Revision a4ae672) },
79
+ year = 2024,
80
+ url = { https://huggingface.co/datasets/pfilonenko/ML_for_TwoSampleTesting },
81
+ doi = { 10.57967/hf/2978 },
82
+ publisher = { Hugging Face }
83
+ }
84
+ ~~~
85
+
86
  # Dataset Simulation
87
 
88
  For this dataset, the full source code (C++) is available [here](https://github.com/pfilonenko/ML_for_TwoSampleTesting/tree/main/dataset/simulation).
 
179
  return 0;
180
  }
181
  ```
182
+ > simulation_for_machine_learning.h
183
+ ```C++
184
+ #ifndef simulation_for_machine_learning_H
185
+ #define simulation_for_machine_learning_H
186
 
187
+ #include"HelpFucntions.h"
188
+
189
+ // Object of the data simulation for training of the proposed ML-method
190
+ class simulation_for_machine_learning{
191
+ private:
192
+ // p-value computation using the Test and Test Statistic (Sn)
193
+ double pvalue(double Sn, HomogeneityTest* Test)
194
+ {
195
+ auto f = Test->F( Sn );
196
+ double pv = 0;
197
+ if( Test->TestType().c_str() == "right" )
198
+ pv = 1.0 - f;
199
+ else
200
+ if( Test->TestType().c_str() == "left" )
201
+ pv = f;
202
+ else // "double"
203
+ pv = 2.0*min( f, 1-f );
204
+ return pv;
205
+ }
206
+
207
+ // Process of simulation
208
+ void Simulation(int iter, vector<HomogeneityTest*> &D, int rank, mt19937boost Gw)
209
+ {
210
+ // ñôîðìèðîâàëè íàçâàíèå ôàéëà äëÿ ñîõðàíåíèÿ
211
+ char file_to_save[512];
212
+ sprintf(file_to_save,".//to_machine_learning_2024//to_machine_learning[rank=%d].csv", rank);
213
+
214
+ // åñëè ýòî ñàìàÿ ïåðâàÿ èòåðàöèÿ, òî ñîõðàíèëè øàïêó ôàéëà
215
+ if( iter == 0 )
216
+ {
217
+ FILE *ou = fopen(file_to_save,"w");
218
+ fprintf(ou, "num;H0/H1;model;n1;n2;perc;real_perc1;real_perc2;");
219
+ for(int i=0; i<D.size(); i++)
220
+ {
221
+ char title_of_test[512];
222
+ D[i]->TitleTest(title_of_test);
223
+ fprintf(ou, "Sn [%s];p-value [%s];", title_of_test, title_of_test);
224
+ }
225
+ fprintf(ou, "\n");
226
+ fclose(ou);
227
+ }
228
+
229
+ // Getting list of the Alternative Hypotheses (H01 - H27)
230
+ vector<int> H;
231
+ int l = 1;
232
+ for(int i=100; i<940; i+=100) // Groups of Alternative Hypotheses (I, II, III, IV, V, VI, VII, VIII, IX)
233
+ {
234
+ for(int j=10; j<40; j+=10) // Alternative Hypotheses in the Group (e.g., H01, H02, H03 into the I and so on)
235
+ //for(int l=1; l<4; l++) // various families of distribution of censoring time F^C(t)
236
+ H.push_back( 1000+i+j+l );
237
+ }
238
+
239
+ // Sample sizes
240
+ vector<int> sample_sizes;
241
+ sample_sizes.push_back( 20 ); // n1 = n2 = 20
242
+ sample_sizes.push_back( 30 ); // n1 = n2 = 30
243
+ sample_sizes.push_back( 50 ); // n1 = n2 = 50
244
+ sample_sizes.push_back( 75 ); // n1 = n2 = 75
245
+ sample_sizes.push_back( 100 ); // n1 = n2 = 100
246
+ sample_sizes.push_back( 150 ); // n1 = n2 = 150
247
+ sample_sizes.push_back( 200 ); // n1 = n2 = 200
248
+ sample_sizes.push_back( 300 ); // n1 = n2 = 300
249
+ sample_sizes.push_back( 500 ); // n1 = n2 = 500
250
+ sample_sizes.push_back( 1000 ); // n1 = n2 = 1000
251
+
252
+ // Simulation (Getting H, Simulation samples, Computation of the test statistics & Save to file)
253
+ for(int i = 0; i<H.size(); i++)
254
+ {
255
+ int Hyp = H[i];
256
+
257
+ if(rank == 0)
258
+ printf("\tH = %d\n",Hyp);
259
+
260
+ for(int per = 0; per<51; per+=10)
261
+ {
262
+ // ---- Getting Hi ----
263
+ AlternativeHypotheses H0_1(Hyp,1,0), H0_2(Hyp,2,0);
264
+ AlternativeHypotheses H1_1(Hyp,1,per), H1_2(Hyp,2,per);
265
+
266
+ for(int jj=0; jj<sample_sizes.size(); jj++)
267
+ {
268
+ int n = sample_sizes[jj];
269
+
270
+ // ---- Simulation samples ----
271
+ //competing hypothesis Í0
272
+ Sample A0(*H0_1.D,n,Gw);
273
+ Sample B0(*H0_1.D,n,Gw);
274
+ if( per > 0 )
275
+ {
276
+ A0.CensoredTypeThird(*H1_1.D,Gw);
277
+ B0.CensoredTypeThird(*H1_1.D,Gw);
278
+ }
279
+
280
+ //competing hypothesis Í1
281
+ Sample A1(*H0_1.D,n,Gw);
282
+ Sample B1(*H0_2.D,n,Gw);
283
+ if( per > 0 )
284
+ {
285
+ A1.CensoredTypeThird(*H1_1.D,Gw);
286
+ B1.CensoredTypeThird(*H1_2.D,Gw);
287
+ }
288
+
289
+ // ---- Computation of the test statistics & Save to file ----
290
+ //Sn and p-value computation under Í0
291
+ FILE *ou = fopen(file_to_save, "a");
292
+ auto perc1 = A0.RealCensoredPercent();
293
+ auto perc2 = B0.RealCensoredPercent();
294
+ fprintf(ou,"%d;", iter);
295
+ fprintf(ou,"H0;");
296
+ fprintf(ou,"%d;", Hyp);
297
+ fprintf(ou,"%d;%d;", n,n);
298
+ fprintf(ou,"%d;%lf;%lf", per, perc1, perc2);
299
+ for(int j=0; j<D.size(); j++)
300
+ {
301
+ auto Sn_H0 = D[j]->CalculateStatistic(A0, B0);
302
+ auto pv_H0 = 0.0; // skip computation (it prepares in ML-framework)
303
+ fprintf(ou, ";%lf;0", Sn_H0);
304
+ }
305
+ fprintf(ou, "\n");
306
+
307
+ //Sn and p-value computation under Í1
308
+ perc1 = A1.RealCensoredPercent();
309
+ perc2 = B1.RealCensoredPercent();
310
+ fprintf(ou,"%d;", iter);
311
+ fprintf(ou,"H1;");
312
+ fprintf(ou,"%d;", Hyp);
313
+ fprintf(ou,"%d;%d;", n,n);
314
+ fprintf(ou,"%d;%lf;%lf", per, perc1, perc2);
315
+ for(int j=0; j<D.size(); j++)
316
+ {
317
+ auto Sn_H1 = D[j]->CalculateStatistic(A1, B1);
318
+ auto pv_H1 = 0.0; // skip computation (it prepares in ML-framework)
319
+ fprintf(ou, ";%lf;0", Sn_H1);
320
+ }
321
+ fprintf(ou, "\n");
322
+ fclose( ou );
323
+ }
324
+ }
325
+ }
326
+ }
327
+
328
+ public:
329
+ // Constructor of the class
330
+ simulation_for_machine_learning(vector<HomogeneityTest*> &D)
331
+ {
332
+ int N = 40000; // number of the Monte-Carlo replications
333
+ #pragma omp parallel for
334
+ for(int k=0; k<N; k++)
335
+ {
336
+ int rank = omp_get_thread_num();
337
+ auto gen = GwMT19937[rank];
338
+
339
+ if(rank == 0)
340
+ printf("\r%d", k);
341
+
342
+ Simulation(k, D, rank, gen);
343
+ }
344
+ }
345
+ };
346
+
347
+ #endif
348
+ ```
349