[1397] | 1 | |
---|
| 2 | /*! |
---|
| 3 | \file |
---|
| 4 | \brief Robust |
---|
| 5 | \author Vasek Smidl |
---|
| 6 | |
---|
| 7 | */ |
---|
| 8 | |
---|
| 9 | #include "estim/arx.h" |
---|
| 10 | #include "robustlib.h" |
---|
| 11 | #include <vector> |
---|
| 12 | #include <iostream> |
---|
| 13 | #include <fstream> |
---|
| 14 | //#include <itpp/itsignal.h> |
---|
| 15 | #include "windows.h" |
---|
| 16 | #include "ddeml.h" |
---|
| 17 | #include "stdio.h" |
---|
| 18 | |
---|
| 19 | //#include "DDEClient.h" |
---|
| 20 | //#include <conio.h> |
---|
| 21 | |
---|
| 22 | |
---|
| 23 | using namespace itpp; |
---|
| 24 | using namespace bdm; |
---|
| 25 | |
---|
| 26 | //const int emlig_size = 2; |
---|
| 27 | //const int utility_constant = 5; |
---|
| 28 | |
---|
| 29 | const int max_model_order = 1; |
---|
| 30 | const double apriorno = 0.01; |
---|
| 31 | const int max_window_size = 121; |
---|
| 32 | |
---|
| 33 | /* |
---|
| 34 | HDDEDATA CALLBACK DdeCallback( |
---|
| 35 | UINT uType, // Transaction type. |
---|
| 36 | UINT uFmt, // Clipboard data format. |
---|
| 37 | HCONV hconv, // Handle to the conversation. |
---|
| 38 | HSZ hsz1, // Handle to a string. |
---|
| 39 | HSZ hsz2, // Handle to a string. |
---|
| 40 | HDDEDATA hdata, // Handle to a global memory object. |
---|
| 41 | DWORD dwData1, // Transaction-specific data. |
---|
| 42 | DWORD dwData2) // Transaction-specific data. |
---|
| 43 | { |
---|
| 44 | return 0; |
---|
| 45 | } |
---|
| 46 | |
---|
| 47 | void DDERequest(DWORD idInst, HCONV hConv, char* szItem) |
---|
| 48 | { |
---|
| 49 | HSZ hszItem = DdeCreateStringHandle(idInst, szItem, 0); |
---|
| 50 | HDDEDATA hData = DdeClientTransaction(NULL,0,hConv,hszItem,CF_TEXT, |
---|
| 51 | XTYP_ADVSTART,TIMEOUT_ASYNC , NULL); //TIMEOUT_ASYNC |
---|
| 52 | if (hData==NULL) |
---|
| 53 | { |
---|
| 54 | printf("Request failed: %s\n", szItem); |
---|
| 55 | } |
---|
| 56 | |
---|
| 57 | if (hData==0) |
---|
| 58 | { |
---|
| 59 | printf("Request failed: %s\n", szItem); |
---|
| 60 | } |
---|
| 61 | } |
---|
| 62 | |
---|
| 63 | DWORD WINAPI ThrdFunc( LPVOID n ) |
---|
| 64 | { |
---|
| 65 | return 0; |
---|
| 66 | } |
---|
| 67 | */ |
---|
| 68 | |
---|
| 69 | class model |
---|
| 70 | { |
---|
| 71 | public: |
---|
| 72 | set<pair<int,int>> ar_components; |
---|
| 73 | |
---|
| 74 | // Best thing would be to inherit the two models from a single souce, this is planned, but now structurally |
---|
| 75 | // problematic. |
---|
| 76 | RARX* my_rarx; //vzmenovane parametre pre triedu model |
---|
| 77 | ARXwin* my_arx; |
---|
| 78 | |
---|
| 79 | bool has_constant; |
---|
| 80 | int window_size; //musi byt vacsia ako pocet krokov ak to nema ovplyvnit |
---|
| 81 | int predicted_channel; |
---|
| 82 | mat* data_matrix; |
---|
| 83 | vec predictions; |
---|
| 84 | char name[80]; |
---|
| 85 | |
---|
| 86 | model(set<pair<int,int>> ar_components, //funkcie treidz model-konstruktor |
---|
| 87 | bool robust, |
---|
| 88 | bool has_constant, |
---|
| 89 | int window_size, |
---|
| 90 | int predicted_channel, |
---|
| 91 | mat* data_matrix) |
---|
| 92 | { |
---|
| 93 | this->ar_components.insert(ar_components.begin(),ar_components.end()); |
---|
| 94 | |
---|
| 95 | strcpy(name,"M"); |
---|
| 96 | |
---|
| 97 | for(set<pair<int,int>>::iterator ar_ref = ar_components.begin();ar_ref!=ar_components.end();ar_ref++) |
---|
| 98 | { |
---|
| 99 | char buffer1[2]; |
---|
| 100 | char buffer2[2]; |
---|
| 101 | itoa((*ar_ref).first,buffer1,10); |
---|
| 102 | itoa((*ar_ref).second,buffer2,10); |
---|
| 103 | |
---|
| 104 | strcat(name,buffer1); |
---|
| 105 | strcat(name,buffer2); |
---|
| 106 | strcat(name,"_"); |
---|
| 107 | } |
---|
| 108 | |
---|
| 109 | this->has_constant = has_constant; |
---|
| 110 | |
---|
| 111 | if(has_constant) |
---|
| 112 | { |
---|
| 113 | strcat(name,"C"); |
---|
| 114 | } |
---|
| 115 | |
---|
| 116 | this->window_size = window_size; |
---|
| 117 | this->predicted_channel = predicted_channel; |
---|
| 118 | this->data_matrix = data_matrix; |
---|
| 119 | |
---|
| 120 | if(robust) |
---|
| 121 | { |
---|
| 122 | strcat(name,"R"); |
---|
| 123 | |
---|
| 124 | if(has_constant) |
---|
| 125 | { |
---|
| 126 | my_rarx = new RARX(ar_components.size()+1,window_size,true,sqrt(2*apriorno),sqrt(2*apriorno),ar_components.size()+4); |
---|
| 127 | my_arx = NULL; |
---|
| 128 | } |
---|
| 129 | else |
---|
| 130 | { |
---|
| 131 | my_rarx = new RARX(ar_components.size(),window_size,false,sqrt(2*apriorno),sqrt(2*apriorno),ar_components.size()+3); |
---|
| 132 | my_arx = NULL; |
---|
| 133 | } |
---|
| 134 | } |
---|
| 135 | else |
---|
| 136 | { |
---|
| 137 | my_rarx = NULL; |
---|
| 138 | my_arx = new ARXwin(); |
---|
| 139 | mat V0; |
---|
| 140 | |
---|
| 141 | if(has_constant) |
---|
| 142 | { |
---|
| 143 | V0 = apriorno * eye(ar_components.size()+2); //aj tu konst |
---|
| 144 | //V0(0,0) = 0; |
---|
| 145 | my_arx->set_constant(true); |
---|
| 146 | } |
---|
| 147 | else |
---|
| 148 | { |
---|
| 149 | |
---|
| 150 | V0 = apriorno * eye(ar_components.size()+1);//menit konstantu |
---|
| 151 | //V0(0,1) = -0.01; |
---|
| 152 | //V0(1,0) = -0.01; |
---|
| 153 | my_arx->set_constant(false); |
---|
| 154 | |
---|
| 155 | } |
---|
| 156 | |
---|
| 157 | my_arx->set_statistics(1, V0, V0.rows()+2); |
---|
| 158 | my_arx->set_parameters(window_size); |
---|
| 159 | my_arx->validate(); |
---|
| 160 | |
---|
| 161 | vec mean = my_arx->posterior().mean(); |
---|
| 162 | cout << mean << endl; |
---|
| 163 | } |
---|
| 164 | } |
---|
| 165 | |
---|
| 166 | void data_update(int time) //vlozime cas a ono vlozi do data_vector podmineky(conditions) a predikce, ktore pouzije do bayes |
---|
| 167 | { |
---|
| 168 | vec data_vector; |
---|
| 169 | for(set<pair<int,int>>::iterator ar_iterator = ar_components.begin();ar_iterator!=ar_components.end();ar_iterator++) |
---|
| 170 | { //ar?iterator ide len od 1 pod 2, alebo niekedy len 1 |
---|
| 171 | data_vector.ins(data_vector.size(),(*data_matrix).get(ar_iterator->first,time-ar_iterator->second)); |
---|
| 172 | // do data vector vlozi pre dany typ regresoru prislusne cisla z data_matrix. Ale ako? preco time-ar_iterator->second? |
---|
| 173 | } |
---|
| 174 | if(my_rarx!=NULL) |
---|
| 175 | { //pre robust priradi az tu do data_vector aj predikciu |
---|
| 176 | data_vector.ins(0,(*data_matrix).get(predicted_channel,time)); |
---|
| 177 | my_rarx->bayes(data_vector); |
---|
| 178 | } |
---|
| 179 | else |
---|
| 180 | { |
---|
| 181 | vec pred_vec;//tu sa predikcia zadava zvlast |
---|
| 182 | pred_vec.ins(0,(*data_matrix).get(predicted_channel,time)); |
---|
| 183 | my_arx->bayes(pred_vec,data_vector); |
---|
| 184 | } |
---|
| 185 | } |
---|
| 186 | |
---|
| 187 | pair<vec,vec> predict(int sample_size, int time, itpp::Laplace_RNG* LapRNG) //nerozumiem, ale vraj to netreba, nepouziva to |
---|
| 188 | { |
---|
| 189 | vec condition_vector; |
---|
| 190 | for(set<pair<int,int>>::iterator ar_iterator = ar_components.begin();ar_iterator!=ar_components.end();ar_iterator++) |
---|
| 191 | { |
---|
| 192 | condition_vector.ins(condition_vector.size(),(*data_matrix).get(ar_iterator->first,time-ar_iterator->second+1)); |
---|
| 193 | } |
---|
| 194 | |
---|
| 195 | if(my_rarx!=NULL) |
---|
| 196 | { |
---|
| 197 | pair<vec,mat> imp_samples = my_rarx->posterior->sample(sample_size,false); |
---|
| 198 | |
---|
| 199 | //cout << imp_samples.first << endl; |
---|
| 200 | |
---|
| 201 | vec sample_prediction; |
---|
| 202 | for(int t = 0;t<sample_size;t++) |
---|
| 203 | { |
---|
| 204 | vec lap_sample = condition_vector; |
---|
| 205 | |
---|
| 206 | if(has_constant) |
---|
| 207 | { |
---|
| 208 | lap_sample.ins(lap_sample.size(),1.0); |
---|
| 209 | } |
---|
| 210 | |
---|
| 211 | lap_sample.ins(lap_sample.size(),(*LapRNG)()); |
---|
| 212 | |
---|
| 213 | sample_prediction.ins(0,lap_sample*imp_samples.second.get_col(t)); |
---|
| 214 | } |
---|
| 215 | |
---|
| 216 | return pair<vec,vec>(imp_samples.first,sample_prediction); |
---|
| 217 | } |
---|
| 218 | else |
---|
| 219 | { |
---|
| 220 | mat samples = my_arx->posterior().sample_mat(sample_size); |
---|
| 221 | |
---|
| 222 | vec sample_prediction; |
---|
| 223 | for(int t = 0;t<sample_size;t++) |
---|
| 224 | { |
---|
| 225 | vec gau_sample = condition_vector; |
---|
| 226 | |
---|
| 227 | if(has_constant) |
---|
| 228 | { |
---|
| 229 | gau_sample.ins(gau_sample.size(),1.0); |
---|
| 230 | } |
---|
| 231 | |
---|
| 232 | gau_sample.ins(gau_sample.size(),randn()); |
---|
| 233 | |
---|
| 234 | sample_prediction.ins(0,gau_sample*samples.get_col(t)); |
---|
| 235 | } |
---|
| 236 | |
---|
| 237 | return pair<vec,vec>(ones(sample_prediction.size()),sample_prediction); |
---|
| 238 | } |
---|
| 239 | |
---|
| 240 | } |
---|
| 241 | |
---|
| 242 | |
---|
| 243 | static set<set<pair<int,int>>> possible_models_recurse(int max_order,int number_of_channels) |
---|
| 244 | { |
---|
| 245 | set<set<pair<int,int>>> created_model_types; |
---|
| 246 | |
---|
| 247 | if(max_order == 1)//ukoncovacia vetva |
---|
| 248 | { |
---|
| 249 | for(int channel = 0;channel<number_of_channels;channel++)//pre AR 1 model vytvori kombinace kanalov v prvom kroku poyadu |
---|
| 250 | { |
---|
| 251 | set<pair<int,int>> returned_type; |
---|
| 252 | returned_type.insert(pair<int,int>(channel,1)); //?? |
---|
| 253 | created_model_types.insert(returned_type); |
---|
| 254 | } |
---|
| 255 | |
---|
| 256 | return created_model_types; |
---|
| 257 | } |
---|
| 258 | else |
---|
| 259 | { |
---|
| 260 | created_model_types = possible_models_recurse(max_order-1,number_of_channels);//tu uz mame ulozene kombinace o jeden krok dozadu //rekuryivne volanie |
---|
| 261 | set<set<pair<int,int>>> returned_types; |
---|
| 262 | |
---|
| 263 | for(set<set<pair<int,int>>>::iterator model_ref = created_model_types.begin();model_ref!=created_model_types.end();model_ref++) |
---|
| 264 | { |
---|
| 265 | |
---|
| 266 | for(int order = 1; order<=max_order; order++) |
---|
| 267 | { |
---|
| 268 | for(int channel = 0;channel<number_of_channels;channel++) |
---|
| 269 | { |
---|
| 270 | set<pair<int,int>> returned_type; |
---|
| 271 | pair<int,int> new_pair = pair<int,int>(channel,order);//?? |
---|
| 272 | if(find((*model_ref).begin(),(*model_ref).end(),new_pair)==(*model_ref).end()) //?? |
---|
| 273 | { |
---|
| 274 | returned_type.insert((*model_ref).begin(),(*model_ref).end()); //co vlozi na zaciatok retuned_type? |
---|
| 275 | returned_type.insert(new_pair); |
---|
| 276 | |
---|
| 277 | |
---|
| 278 | returned_types.insert(returned_type); |
---|
| 279 | } |
---|
| 280 | } |
---|
| 281 | } |
---|
| 282 | } |
---|
| 283 | |
---|
| 284 | created_model_types.insert(returned_types.begin(),returned_types.end()); |
---|
| 285 | |
---|
| 286 | return created_model_types; |
---|
| 287 | } |
---|
| 288 | } |
---|
| 289 | }; |
---|
| 290 | |
---|
| 291 | |
---|
| 292 | |
---|
| 293 | |
---|
| 294 | int main ( int argc, char* argv[] ) |
---|
| 295 | { |
---|
| 296 | |
---|
| 297 | // EXPERIMENT: A moving window estimation and prediction of RARX is tested on data generated from |
---|
| 298 | // y_t=0.95*y_(t-1)+0.05*y_(t-2)+0.2*e_t, where e_t is normally, student(4) and cauchy distributed. It |
---|
| 299 | // can be compared to the classical setup. |
---|
| 300 | |
---|
| 301 | itpp::Laplace_RNG LapRNG = Laplace_RNG(); |
---|
| 302 | |
---|
| 303 | vector<vector<string>> strings; |
---|
| 304 | |
---|
| 305 | char* file_string = "C:\\results\\cauchy"; // "C:\\dataADClosePercDiff"; // |
---|
| 306 | |
---|
| 307 | char dfstring[80]; |
---|
| 308 | strcpy(dfstring,file_string); |
---|
| 309 | strcat(dfstring,".txt"); |
---|
| 310 | |
---|
| 311 | |
---|
| 312 | mat data_matrix; |
---|
| 313 | ifstream myfile(dfstring); |
---|
| 314 | if (myfile.is_open()) |
---|
| 315 | { |
---|
| 316 | string line; |
---|
| 317 | while(getline(myfile,line)) |
---|
| 318 | { |
---|
| 319 | vec data_vector; |
---|
| 320 | while(line.find(',') != string::npos) //zmenil som ciarku za medzeru |
---|
| 321 | { |
---|
| 322 | //line.erase(0,1); // toto som sem pridal |
---|
| 323 | int loc2 = line.find('\n'); |
---|
| 324 | int loc = line.find(','); |
---|
| 325 | data_vector.ins(data_vector.size(),atof(line.substr(0,loc).c_str())); |
---|
| 326 | line.erase(0,loc+1); |
---|
| 327 | } |
---|
| 328 | |
---|
| 329 | data_matrix.ins_row(data_matrix.rows(),data_vector); |
---|
| 330 | } |
---|
| 331 | |
---|
| 332 | myfile.close(); |
---|
| 333 | } |
---|
| 334 | else |
---|
| 335 | { |
---|
| 336 | cout << "Can't open data file!" << endl; |
---|
| 337 | } |
---|
| 338 | |
---|
| 339 | //konec nacitavania dat |
---|
| 340 | set<set<pair<int,int>>> model_types = model::possible_models_recurse(max_model_order,data_matrix.rows()); //volanie funkce kde robi kombinace modelov |
---|
| 341 | //to priradime do model_types, data_matrix.row urcuje pocet kanalov dat |
---|
| 342 | vector<model*> models; |
---|
| 343 | for(set<set<pair<int,int>>>::iterator model_type = model_types.begin();model_type!=model_types.end();model_type++) |
---|
| 344 | {// prechadza rozne typy kanalov, a poctu regresorov |
---|
| 345 | for(int window_size = max_window_size-1;window_size < max_window_size;window_size++) |
---|
| 346 | { |
---|
| 347 | //models.push_back(new model((*model_type),true,true,window_size,0,&data_matrix)); // to su len konstruktory, len inicializujeme rozne typy |
---|
| 348 | //models.push_back(new model((*model_type),false,true,window_size,0,&data_matrix)); |
---|
| 349 | models.push_back(new model((*model_type),true,false,window_size,0,&data_matrix)); |
---|
| 350 | models.push_back(new model((*model_type),false,false,window_size,0,&data_matrix)); |
---|
| 351 | } |
---|
| 352 | |
---|
| 353 | //set<pair<int,int>> empty_list; |
---|
| 354 | //models.push_back(new model(empty_list,false,true,100,0,&data_matrix)); |
---|
| 355 | } |
---|
| 356 | |
---|
| 357 | mat result_lognc; |
---|
| 358 | // mat result_preds; |
---|
| 359 | |
---|
| 360 | ofstream myfilew; |
---|
| 361 | //char fstring[80]; |
---|
| 362 | //strcpy(fstring,file_string); |
---|
| 363 | //strcat(fstring,"lognc.txt"); |
---|
| 364 | //strcat(fstring,"preds.txt"); |
---|
| 365 | |
---|
| 366 | for(int time = max_model_order;time<data_matrix.cols();time++) //time<data_matrix.cols() |
---|
| 367 | { |
---|
| 368 | if(time==max_window_size) |
---|
| 369 | { |
---|
| 370 | exit(1); |
---|
| 371 | } |
---|
| 372 | |
---|
| 373 | //pocet stlpcov data_matrix je pocet casovych krokov |
---|
| 374 | vec cur_res_lognc; |
---|
| 375 | // vec preds; |
---|
| 376 | vector<string> nazvy; |
---|
| 377 | for(vector<model*>::iterator model_ref = models.begin();model_ref!=models.end();model_ref++) |
---|
| 378 | {//posuvam s apo models, co je pole modelov urobene o cyklus vyssie. Teda som v case time a robim to tam pre vsetky typy modelov, kombinace regresorov |
---|
| 379 | (*model_ref)->data_update(time); //pozret sa preco je toto tu nutne |
---|
| 380 | |
---|
| 381 | cout << "Updated." << endl; |
---|
| 382 | //if (time = max_model_order) nazvy.push_back(models.model_ref]);// ako by som mohol dostat nazov modelu? |
---|
| 383 | |
---|
| 384 | if((*model_ref)->my_rarx!=NULL) //vklada normalizacni faktor do cur_res_lognc |
---|
| 385 | { |
---|
| 386 | cout << "Maxlik vertex:" << (*model_ref)->my_rarx->posterior->minimal_vertex->get_coordinates() << endl; |
---|
| 387 | cur_res_lognc.ins(cur_res_lognc.size(),(*model_ref)->my_rarx->posterior->_ll()); |
---|
| 388 | } |
---|
| 389 | else |
---|
| 390 | { |
---|
| 391 | cur_res_lognc.ins(cur_res_lognc.size(),(*model_ref)->my_arx->_ll()); |
---|
| 392 | } |
---|
| 393 | |
---|
| 394 | if(time == max_window_size-1) |
---|
| 395 | { |
---|
| 396 | //*********************** |
---|
| 397 | int sample_size = 100000; |
---|
| 398 | //*********************** |
---|
| 399 | |
---|
| 400 | pair<vec,mat> samples; |
---|
| 401 | if((*model_ref)->my_arx!=NULL) |
---|
| 402 | { |
---|
| 403 | mat samp_mat = (*model_ref)->my_arx->posterior().sample_mat(sample_size); |
---|
| 404 | samples = pair<vec,mat>(ones(samp_mat.cols()),samp_mat); |
---|
| 405 | } |
---|
| 406 | else |
---|
| 407 | { |
---|
| 408 | samples = (*model_ref)->my_rarx->posterior->sample(sample_size,true); |
---|
| 409 | } |
---|
| 410 | |
---|
| 411 | char fstring[80]; |
---|
| 412 | strcpy(fstring,file_string); |
---|
| 413 | strcat(fstring,(*model_ref)->name); |
---|
| 414 | strcat(fstring,".txt"); |
---|
| 415 | |
---|
| 416 | //cout << samples.first << endl; |
---|
| 417 | |
---|
| 418 | myfilew.open(fstring,ios::app); |
---|
| 419 | |
---|
| 420 | /* |
---|
| 421 | for(int i = 0;i<samples.first.size();i++) |
---|
| 422 | { |
---|
| 423 | myfilew << samples.first.get(i) << ","; |
---|
| 424 | } |
---|
| 425 | myfilew << endl; |
---|
| 426 | */ |
---|
| 427 | |
---|
| 428 | for(int j = 0;j<samples.second.rows()+1;j++) |
---|
| 429 | { |
---|
| 430 | for(int i = 0;i<samples.second.cols();i++) |
---|
| 431 | { |
---|
| 432 | if(j!=samples.second.rows()) |
---|
| 433 | { |
---|
| 434 | myfilew << samples.second.get(j,i) << ","; |
---|
| 435 | } |
---|
| 436 | /* |
---|
| 437 | else |
---|
| 438 | { |
---|
| 439 | myfilew << "0,"; |
---|
| 440 | } |
---|
| 441 | */ |
---|
| 442 | } |
---|
| 443 | myfilew << endl; |
---|
| 444 | } |
---|
| 445 | |
---|
| 446 | cout << "*************************************" << endl; |
---|
| 447 | |
---|
| 448 | myfilew.close(); |
---|
| 449 | } |
---|
| 450 | |
---|
| 451 | /* // PREDICTIONS |
---|
| 452 | pair<vec,vec> predictions = (*model_ref)->predict(500,time,&LapRNG); |
---|
| 453 | |
---|
| 454 | cout << predictions.first << endl << predictions.second << endl; |
---|
| 455 | |
---|
| 456 | double avg_prediction = (predictions.first*predictions.second)/(predictions.first*ones(predictions.first.size())); |
---|
| 457 | |
---|
| 458 | (*model_ref)->predictions.ins((*model_ref)->predictions.size(),avg_prediction); |
---|
| 459 | |
---|
| 460 | myfilew.open(fstring,ios::app); |
---|
| 461 | myfilew << avg_prediction << ","; |
---|
| 462 | myfilew.close(); |
---|
| 463 | */ |
---|
| 464 | |
---|
| 465 | //preds.ins(0,data_matrix.get(0,time+1)); |
---|
| 466 | } |
---|
| 467 | |
---|
| 468 | |
---|
| 469 | /* // REAL PRICE |
---|
| 470 | myfilew.open(fstring,ios::app); |
---|
| 471 | myfilew << data_matrix.get(0,time+1) << endl; |
---|
| 472 | myfilew.close(); |
---|
| 473 | */ |
---|
| 474 | |
---|
| 475 | result_lognc.ins_col(result_lognc.cols(),cur_res_lognc); |
---|
| 476 | // result_preds.ins_col(result_preds.cols(),preds); |
---|
| 477 | |
---|
| 478 | // cout << "Updated." << endl; |
---|
| 479 | |
---|
| 480 | /* |
---|
| 481 | myfilew.open(fstring,ios::app); |
---|
| 482 | |
---|
| 483 | // myfile << my_rarx->posterior->minimal_vertex->get_coordinates()[0]; |
---|
| 484 | |
---|
| 485 | if(time == max_model_order) |
---|
| 486 | { |
---|
| 487 | for(int i = 0;i<cur_res_lognc.size();i++) |
---|
| 488 | { |
---|
| 489 | for(set<pair<int,int>>::iterator ar_ref = models[i]->ar_components.begin();ar_ref != models[i]->ar_components.end();ar_ref++) |
---|
| 490 | { |
---|
| 491 | myfilew << (*ar_ref).second << (*ar_ref).first; |
---|
| 492 | } |
---|
| 493 | |
---|
| 494 | myfilew << "."; |
---|
| 495 | |
---|
| 496 | if(models[i]->my_arx == NULL) |
---|
| 497 | { |
---|
| 498 | myfilew << "1"; |
---|
| 499 | } |
---|
| 500 | else |
---|
| 501 | { |
---|
| 502 | myfilew << "0"; |
---|
| 503 | } |
---|
| 504 | |
---|
| 505 | if(models[i]->has_constant) |
---|
| 506 | { |
---|
| 507 | myfilew << "1"; |
---|
| 508 | } |
---|
| 509 | else |
---|
| 510 | { |
---|
| 511 | myfilew << "0"; |
---|
| 512 | } |
---|
| 513 | |
---|
| 514 | myfilew << ","; |
---|
| 515 | } |
---|
| 516 | |
---|
| 517 | myfilew << endl; |
---|
| 518 | } |
---|
| 519 | |
---|
| 520 | |
---|
| 521 | // for(int i = 0;i<cur_res_lognc.size();i++) |
---|
| 522 | // { |
---|
| 523 | // myfilew << cur_res_lognc[i] << ' ';//zmenil som ciarku ze medzeru |
---|
| 524 | // } |
---|
| 525 | |
---|
| 526 | |
---|
| 527 | myfilew << endl; |
---|
| 528 | |
---|
| 529 | myfilew.close(); |
---|
| 530 | */ |
---|
| 531 | |
---|
| 532 | } |
---|
| 533 | |
---|
| 534 | |
---|
| 535 | |
---|
| 536 | // EXPERIMENT: One step ahead price prediction. Comparison of classical and robust model using optimal trading |
---|
| 537 | // with maximization of logarithm of one-step ahead wealth. |
---|
| 538 | |
---|
| 539 | |
---|
| 540 | |
---|
| 541 | /* |
---|
| 542 | cout << "One experiment finished." << endl; |
---|
| 543 | |
---|
| 544 | ofstream myfile; |
---|
| 545 | myfile.open("c:\\robust_ar1.txt",ios::app); |
---|
| 546 | myfile << endl; |
---|
| 547 | myfile.close(); |
---|
| 548 | |
---|
| 549 | myfile.open("c:\\robust_ar2.txt",ios::app); |
---|
| 550 | myfile << endl; |
---|
| 551 | myfile.close();*/ |
---|
| 552 | |
---|
| 553 | |
---|
| 554 | //emlig* emlig1 = new emlig(emlig_size); |
---|
| 555 | |
---|
| 556 | //emlig1->step_me(0); |
---|
| 557 | //emlig* emlig2 = new emlig(emlig_size); |
---|
| 558 | |
---|
| 559 | /* |
---|
| 560 | emlig1->set_correction_factors(4); |
---|
| 561 | |
---|
| 562 | for(int j = 0;j<emlig1->correction_factors.size();j++) |
---|
| 563 | { |
---|
| 564 | for(set<my_ivec>::iterator vec_ref = emlig1->correction_factors[j].begin();vec_ref!=emlig1->correction_factors[j].end();vec_ref++) |
---|
| 565 | { |
---|
| 566 | cout << j << " "; |
---|
| 567 | |
---|
| 568 | for(int i=0;i<(*vec_ref).size();i++) |
---|
| 569 | { |
---|
| 570 | cout << (*vec_ref)[i]; |
---|
| 571 | } |
---|
| 572 | |
---|
| 573 | cout << endl; |
---|
| 574 | } |
---|
| 575 | }*/ |
---|
| 576 | |
---|
| 577 | /* |
---|
| 578 | vec condition5 = "1.0 1.0 1.01";//"-0.3 1.7 1.5"; |
---|
| 579 | |
---|
| 580 | emlig1->add_condition(condition5); |
---|
| 581 | //emlig1->step_me(0); |
---|
| 582 | |
---|
| 583 | |
---|
| 584 | vec condition1a = "-1.0 1.02 0.5"; |
---|
| 585 | //vec condition1b = "1.0 1.0 1.01"; |
---|
| 586 | emlig1->add_condition(condition1a); |
---|
| 587 | //emlig2->add_condition(condition1b); |
---|
| 588 | |
---|
| 589 | vec condition2a = "-0.3 1.7 1.5"; |
---|
| 590 | //vec condition2b = "-1.0 1.0 1.0"; |
---|
| 591 | emlig1->add_condition(condition2a); |
---|
| 592 | //emlig2->add_condition(condition2b); |
---|
| 593 | |
---|
| 594 | vec condition3a = "0.5 -1.01 1.0"; |
---|
| 595 | //vec condition3b = "0.5 -1.01 1.0"; |
---|
| 596 | |
---|
| 597 | emlig1->add_condition(condition3a); |
---|
| 598 | //emlig2->add_condition(condition3b); |
---|
| 599 | |
---|
| 600 | vec condition4a = "-0.5 -1.0 1.0"; |
---|
| 601 | //vec condition4b = "-0.5 -1.0 1.0"; |
---|
| 602 | |
---|
| 603 | emlig1->add_condition(condition4a); |
---|
| 604 | //cout << "************************************************" << endl; |
---|
| 605 | //emlig2->add_condition(condition4b); |
---|
| 606 | //cout << "************************************************" << endl; |
---|
| 607 | |
---|
| 608 | //cout << emlig1->minimal_vertex->get_coordinates(); |
---|
| 609 | |
---|
| 610 | //emlig1->remove_condition(condition3a); |
---|
| 611 | //emlig1->step_me(0); |
---|
| 612 | //emlig1->remove_condition(condition2a); |
---|
| 613 | //emlig1->remove_condition(condition1a); |
---|
| 614 | //emlig1->remove_condition(condition5); |
---|
| 615 | |
---|
| 616 | |
---|
| 617 | //emlig1->step_me(0); |
---|
| 618 | //emlig2->step_me(0); |
---|
| 619 | |
---|
| 620 | |
---|
| 621 | // DA SE POUZIT PRO VYPIS DO SOUBORU |
---|
| 622 | // emlig1->step_me(0); |
---|
| 623 | |
---|
| 624 | //emlig1->remove_condition(condition1); |
---|
| 625 | |
---|
| 626 | |
---|
| 627 | |
---|
| 628 | |
---|
| 629 | |
---|
| 630 | /* |
---|
| 631 | for(int i = 0;i<100;i++) |
---|
| 632 | { |
---|
| 633 | cout << endl << "Step:" << i << endl; |
---|
| 634 | |
---|
| 635 | double condition[emlig_size+1]; |
---|
| 636 | |
---|
| 637 | for(int k = 0;k<=emlig_size;k++) |
---|
| 638 | { |
---|
| 639 | condition[k] = (rand()-RAND_MAX/2)/1000.0; |
---|
| 640 | } |
---|
| 641 | |
---|
| 642 | |
---|
| 643 | vec* condition_vec = new vec(condition,emlig_size+1); |
---|
| 644 | emlig1->add_condition(*condition_vec); |
---|
| 645 | |
---|
| 646 | /* |
---|
| 647 | for(polyhedron* toprow_ref = emlig1->statistic.rows[emlig_size]; toprow_ref != emlig1->statistic.end_poly; toprow_ref = toprow_ref->next_poly) |
---|
| 648 | { |
---|
| 649 | cout << ((toprow*)toprow_ref)->probability << endl; |
---|
| 650 | } |
---|
| 651 | */ |
---|
| 652 | /* |
---|
| 653 | cout << emlig1->statistic_rowsize(emlig_size) << endl << endl; |
---|
| 654 | |
---|
| 655 | /* |
---|
| 656 | if(i-emlig1->number_of_parameters >= 0) |
---|
| 657 | { |
---|
| 658 | pause(30); |
---|
| 659 | } |
---|
| 660 | */ |
---|
| 661 | |
---|
| 662 | // emlig1->step_me(i); |
---|
| 663 | |
---|
| 664 | /* |
---|
| 665 | vector<int> sizevector; |
---|
| 666 | for(int s = 0;s<=emlig1->number_of_parameters;s++) |
---|
| 667 | { |
---|
| 668 | sizevector.push_back(emlig1->statistic_rowsize(s)); |
---|
| 669 | } |
---|
| 670 | */ |
---|
| 671 | //} |
---|
| 672 | |
---|
| 673 | |
---|
| 674 | |
---|
| 675 | |
---|
| 676 | /* |
---|
| 677 | emlig1->step_me(1); |
---|
| 678 | |
---|
| 679 | vec condition = "2.0 0.0 1.0"; |
---|
| 680 | |
---|
| 681 | emlig1->add_condition(condition); |
---|
| 682 | |
---|
| 683 | vector<int> sizevector; |
---|
| 684 | for(int s = 0;s<=emlig1->number_of_parameters;s++) |
---|
| 685 | { |
---|
| 686 | sizevector.push_back(emlig1->statistic_rowsize(s)); |
---|
| 687 | } |
---|
| 688 | |
---|
| 689 | emlig1->step_me(2); |
---|
| 690 | |
---|
| 691 | condition = "2.0 1.0 0.0"; |
---|
| 692 | |
---|
| 693 | emlig1->add_condition(condition); |
---|
| 694 | |
---|
| 695 | sizevector.clear(); |
---|
| 696 | for(int s = 0;s<=emlig1->number_of_parameters;s++) |
---|
| 697 | { |
---|
| 698 | sizevector.push_back(emlig1->statistic_rowsize(s)); |
---|
| 699 | } |
---|
| 700 | */ |
---|
| 701 | |
---|
| 702 | return 0; |
---|
| 703 | } |
---|
| 704 | |
---|
| 705 | |
---|