#include <getopt.h>#include <fcntl.h>#include <errno.h>#include <err.h>#include <cstdio>#include <cstdlib>#include <cstring>#include <cmath>#include <iostream>#include <fstream>#include <string>#include <boost/timer.hpp>#include "svm.h"#include "LibsvmFileLoader.h"#include "SVClustering.h"#include "CCLSVClustering.h"#include "GKWGenerator.h"#include "KernelMachines.h"#include "L1Distance.h"#include "L2Distance.h"Include dependency graph for svc.c:
Go to the source code of this file.
Classes | |
| struct | svc_options |
| The structure encapsulating the SVC runing options. More... | |
Defines | |
| #define | DEFAULT_ALG 0 |
| #define | DEFAULT_C 1 |
| #define | DEFAULT_EDIST 2 |
| #define | DEFAULT_KERNEL -1 |
| #define | DEFAULT_KW -1 |
| Support Vector Clustering Test and Benchmark. | |
| #define | DEFAULT_L_OFFSET 1 |
| #define | DEFAULT_RUNS 100000 |
| #define | DEFAULT_SOFTENING 1 |
| #define | DEFAULT_VALID_THRESHOLD 2 |
Enumerations | |
| enum | { CCL = 0, CG = 1 } |
| The enumeration for Cluster Labeling algorithm types. More... | |
| enum | { L1 = 1, L2 = 2 } |
| The enumeration for Euclidean Distance types. More... | |
| enum | { BSV_CLASSIC = 0, BSV_ENHANCED = 1 } |
| The enumeration for the BSV clustering policy. More... | |
Functions | |
| void | computeEvaluatingMeasreusPerClass (vector< int > tp, vector< int > fp, vector< int > tn, vector< int > fn, vector< double > &contamination, vector< double > &specificity, vector< double > &precision, vector< double > &recall, vector< double > &f1, double &ma, double &accuracy) |
| For all classes computes the precision/recall/f1 values. | |
| void | countRightAndWrongClassificationsPerClass (damina::SVClustering *clust, vector< int > &tp, vector< int > &fp, vector< int > &tn, vector< int > &fn, int nrClasses, int classOffset, vector< unsigned long int > &ppc, unsigned long int valid_t) |
| For all classes computes the number of true/false positives and true/false negatives. | |
| int | main (int argc, char *argv[]) |
| The main flow for SVC benchmark. | |
| svc_options * | parseCommandLineArgs (int argc, char **argv) |
| Parse the command line arguments and return the options for the SVC running. | |
| void | printAllPoints (damina::SVClustering *clust, unsigned long int classOffset, char *filename) |
| Print all points with their original classes and cluster assigned. | |
| void | printUsedCommandLine (int argc, char *argv[]) |
| Print the whole command line used to invoke the program. | |
| void | usage (char **argv) |
| Print the usage on the standard error. | |
| #define DEFAULT_ALG 0 |
| #define DEFAULT_C 1 |
| #define DEFAULT_EDIST 2 |
| #define DEFAULT_KERNEL -1 |
| #define DEFAULT_KW -1 |
Support Vector Clustering Test and Benchmark.
Definition at line 36 of file svc.c.
Referenced by parseCommandLineArgs().
| #define DEFAULT_L_OFFSET 1 |
| #define DEFAULT_RUNS 100000 |
| #define DEFAULT_SOFTENING 1 |
| #define DEFAULT_VALID_THRESHOLD 2 |
| anonymous enum |
| anonymous enum |
| anonymous enum |
The enumeration for the BSV clustering policy.
Definition at line 72 of file svc.c.
00072 {BSV_CLASSIC = 0, BSV_ENHANCED = 1};
| void computeEvaluatingMeasreusPerClass | ( | vector< int > | tp, | |
| vector< int > | fp, | |||
| vector< int > | tn, | |||
| vector< int > | fn, | |||
| vector< double > & | contamination, | |||
| vector< double > & | specificity, | |||
| vector< double > & | precision, | |||
| vector< double > & | recall, | |||
| vector< double > & | f1, | |||
| double & | ma, | |||
| double & | accuracy | |||
| ) |
For all classes computes the precision/recall/f1 values.
| tp | The true positives vector for all classes | |
| fp | The flase positives vector for all classes | |
| tn | The true negatives vector for all classes | |
| fp | The flase negatives vector for all classes | |
| contamination | The contamination value for all classes | |
| precision | The precision value for all classes | |
| recall | The recall value for all classes | |
| f1 | The f1 value for all classes |
Definition at line 396 of file svc.c.
Referenced by main().
00398 { 00399 00400 unsigned long int nrClasses = tp.size(); 00401 int datasedDim = tp[0] + fp[0] + tn[0] + fn[0]; 00402 00403 precision.clear(); 00404 recall.clear(); 00405 f1.clear(); 00406 00407 contamination.resize(nrClasses); 00408 precision.resize(nrClasses); 00409 recall.resize(nrClasses); 00410 f1.resize(nrClasses); 00411 specificity.resize(nrClasses); 00412 00413 ma = 0; 00414 accuracy = 0; 00415 00416 for (unsigned long int i = 0; i < nrClasses ; i++) { 00417 accuracy += tp[i]; 00418 precision[i] = ((double)tp[i] / (double)(tp[i] + fp[i])) * 100; 00419 recall[i] = ((double)tp[i] / (double)(tp[i] + fn[i])) * 100; 00420 specificity[i] = ((double)tn[i] / (double)(tn[i] + fp[i])) * 100; 00421 contamination[i] = (double)fp[i] / (double)tp[i]; 00422 if (precision[i] + recall[i] == 0) { 00423 f1[i] = -1; 00424 } 00425 else { 00426 f1[i] = (2 * (precision[i] * recall[i])) / (precision[i] + recall[i]); 00427 ma += f1[i]; 00428 } 00429 00430 } 00431 00432 ma = ma / (double) nrClasses; 00433 accuracy = accuracy / (double)datasedDim * 100; 00434 }
Here is the caller graph for this function:
| void countRightAndWrongClassificationsPerClass | ( | damina::SVClustering * | clust, | |
| vector< int > & | tp, | |||
| vector< int > & | fp, | |||
| vector< int > & | tn, | |||
| vector< int > & | fn, | |||
| int | nrClasses, | |||
| int | classOffset, | |||
| vector< unsigned long int > & | ppc, | |||
| unsigned long int | valid_t | |||
| ) |
For all classes computes the number of true/false positives and true/false negatives.
| clust | The involved Support Vector Clustering engine | |
| tp | The true positives vector for all classes | |
| fp | The flase positives vector for all classes | |
| tn | The true negatives vector for all classes | |
| fp | The flase negatives vector for all classes | |
| nrClasses | The number of the original classes | |
| classOffset | The label of the first class in the original classification |
Definition at line 349 of file svc.c.
References damina::SVClustering::getClustersAssignment(), damina::SVClustering::getOriginalClasses(), and damina::SVClustering::getProblem().
Referenced by main().
00349 { 00350 00351 std::vector<int> labels = clust->getClustersAssignment(); // clustering result 00352 std::vector<int> orig_c = clust->getOriginalClasses(classOffset); // original classification 00353 00354 int datasetDimension = clust->getProblem()->l; 00355 00356 tp.clear(); 00357 fp.clear(); 00358 tn.clear(); 00359 fn.clear(); 00360 00361 tp.resize(nrClasses); 00362 fp.resize(nrClasses); 00363 tn.resize(nrClasses); 00364 fn.resize(nrClasses); 00365 00366 for (unsigned long int i = 0; i < labels.size(); i++) { 00367 if (orig_c[i] >= nrClasses) continue; 00368 00369 if ((ppc[labels[i]] >= valid_t) && (labels[i] == orig_c[i])) { 00370 tp[orig_c[i]]++; 00371 } 00372 else { 00373 fp[labels[i]]++; 00374 fn[orig_c[i]]++; 00375 } 00376 } 00377 00378 for (int i = 0; i < nrClasses; i++) { 00379 tn[i] = datasetDimension - tp[i] - fp[i] - fn[i]; 00380 } 00381 }
Here is the call graph for this function:
Here is the caller graph for this function:
| int main | ( | int | argc, | |
| char * | argv[] | |||
| ) |
The main flow for SVC benchmark.
Load data file in LibSVM format
New SVC instance
Change kernel type, in requested
Kernel generator instance
Change Euclidean Distance, if requested
Perform clustering
Definition at line 97 of file svc.c.
References svc_options::alg_type, svc_options::bsv, svc_options::C, CCL, CG, svc_options::clusters, computeEvaluatingMeasreusPerClass(), countRightAndWrongClassificationsPerClass(), DEFAULT_KERNEL, svc_options::detfile, svc_options::e_dist, svc_options::filename, svc_options::kernel, L1, L2, svc_options::l_offset, parseCommandLineArgs(), printAllPoints(), printUsedCommandLine(), svc_options::q, svc_options::runs, svc_options::softening, usage(), and svc_options::valid_threshold.
00097 { 00098 00099 using namespace std; 00100 using namespace damina; 00101 using namespace boost; 00102 00103 struct svc_options *opt; 00104 00105 opt = parseCommandLineArgs(argc, argv); // parse command line to get parameters 00106 if (opt == NULL) { // parsing fails 00107 usage(argv); // show usage 00108 exit(1); 00109 } 00110 00114 LibsvmFileLoader *fl; 00115 string datafile(opt->filename); 00116 struct svm_problem *p = NULL; 00117 cout << "==================================" << endl; 00118 cout << "Loading file" << endl; 00119 cout << "==================================" << endl; 00120 fl = new LibsvmFileLoader(datafile); 00121 timer tfl; 00122 p = fl->load(false); 00123 cout << "File loaded in " << tfl.elapsed() << " secs" << endl; 00124 cout << "Number of objects: " << p->l << endl << endl; 00125 delete fl; 00126 00127 00131 SVClustering *svc; // Support Vector Clustering Engine 00132 switch (opt->alg_type) { 00133 case CG: svc = new SVClustering(opt->C, p); break; // classic one 00134 case CCL: svc = new CCLSVClustering(opt->C, p); break; // with CCL 00135 default: 00136 svc = new SVClustering(opt->C, p); break; // default is classic 00137 } 00138 00139 if (opt->bsv) { 00140 svc->useSpecialClusteringPolicyForBSV(); 00141 } 00142 else { 00143 svc->useClassicClusteringPolicyForBSV(); 00144 } 00145 00149 if (opt->kernel != DEFAULT_KERNEL) { 00150 svc->setKernelType(opt->kernel); 00151 } 00152 00156 GKWGenerator *kwg; 00157 if (opt->q == -1) { // no kernel width value provided on command line 00158 kwg = new GKWGenerator(svc, 0.000000001); 00159 } 00160 else { // kernel width value provided on command line 00161 kwg = new GKWGenerator(svc, 0.000000001, opt->q); 00162 } 00163 00164 kwg->setGrowthSoftening(opt->softening); 00165 00169 switch (opt->e_dist) { 00170 case L1: { svc->setEuclideanDistance(new L1Distance()); kwg->setEuclideanDistance(new L1Distance()); } break; 00171 case L2: { svc->setEuclideanDistance(new L2Distance()); kwg->setEuclideanDistance(new L2Distance()); } break; 00172 default: ; 00173 } 00174 00175 bool target = false; 00176 unsigned long int invalid = 0, runs = 0; 00177 double q, prev_q = -1, prev_C = -1; 00178 00179 unsigned long int nrClusters, 00180 // nrNonSingletonClusters; 00181 nrValidClusters; 00182 00183 timer partial, // timer for intermediate steps 00184 total; // timer for overall time measurement 00185 00186 vector<unsigned long int> ppc; 00187 vector<int> tp, fp, tn, fn; 00188 vector<double> precision, recall, f1, contamination, specificity; 00189 double ma, accuracy, overall_time = 0; 00190 00191 while (true) { 00192 prev_q = q; 00193 prev_C = svc->getSoftConstraintEstimate(); 00194 00195 runs++; 00196 if (runs > opt->runs) break; 00197 00198 q = kwg->getNextKernelWidthValue(); 00199 if (q == -1) { 00200 cout << "No more valid kernel width values" << endl << endl; 00201 break; 00202 } 00203 00204 if (opt->runs > 1) { 00205 cout << "==================================" << endl; 00206 cout << " Finding Kernel Width # " << runs << endl; 00207 cout << "==================================" << endl; 00208 } 00209 00210 svc->setKernelWidth(q); 00211 00215 total.restart(); // start total time counter 00216 partial.restart(); // start domain description process time counter 00217 00218 svc->learn(); // domain description process 00219 cout << "radius = " << svc->getSphereRadius() << endl; 00220 cout << endl << "Domain description time taken: "; 00221 cout << partial.elapsed() << " seconds " << endl; // stop domain description process time counter 00222 00223 partial.restart(); // start clustering process time counter 00224 00225 svc->clusterize(); // cluster labeling process 00226 cout << "Clustering time taken: "; 00227 cout << partial.elapsed() << " seconds " << endl; // stop clustering process time counter 00228 00229 cout << "Total time taken: "; 00230 overall_time += total.elapsed(); 00231 cout << total.elapsed() << " seconds " << endl << endl; // stop total time counter 00232 00233 00234 nrClusters = svc->getNumberOfClusters(); 00235 // nrNonSingletonClusters = svc->getNumberOfNonSingletonClusters(); 00236 nrValidClusters = svc->getNumberOfValidClusters(opt->valid_threshold); 00237 00238 cout << "Soft Constraint Estimate: " << svc->getSoftConstraintEstimate() << endl; 00239 cout << "Current Soft Constraint: " << svc->getSoftConstraint() << endl; 00240 cout << "Kernel Width: " << q << endl; 00241 cout << "Number of clusters: " << nrClusters << endl; 00242 cout << "Number of valid clusters: " << nrValidClusters << endl; 00243 cout << "Number of non valid clusters (cardinality < " << opt->valid_threshold << "): " << nrClusters - nrValidClusters << endl; 00244 00245 cout << "Points per cluster: " << endl; 00246 00247 ppc.clear(); 00248 ppc = svc->getPointPerCluster(); 00249 00250 for (unsigned long int i = 0; i < ppc.size(); i++) { 00251 if (ppc[i] >= opt->valid_threshold) { 00252 cout << "\tCluster " << i << ": " << ppc[i] << endl; 00253 } 00254 } 00255 cout << endl << endl; 00256 00257 if (opt->detfile != NULL) { 00258 printAllPoints(svc, opt->l_offset, opt->detfile); 00259 } 00260 00261 countRightAndWrongClassificationsPerClass(svc, tp, fp, tn, fn, opt->clusters, opt->l_offset, ppc, opt->valid_threshold); 00262 computeEvaluatingMeasreusPerClass(tp, fp, tn, fn, contamination, specificity, precision, recall, f1, ma, accuracy); 00263 00264 for (unsigned long int i = 0; i < tp.size(); i++) { 00265 cout << "Class " << i << endl << "\tTP: " << tp[i] << "\tFP: " << fp[i] << endl << "\tFN: " << fn[i] << "\tTN: " << tn[i] << endl << endl; 00266 cout << "Precision: " << precision[i] << " - Recall: " << recall[i] << " - F1: " << f1[i] << endl; 00267 cout << "Completeness: " << recall[i] << " - Contamination: " << contamination[i] << endl; 00268 cout << "Sensitivity: " << recall[i] << " - Specificity: " << specificity[i] << endl << endl; 00269 } 00270 00271 if (opt->clusters > 2) { 00272 cout << "Macroaveraging: " << ma << endl; 00273 } 00274 00275 cout << "Accuracy: " << accuracy << endl << endl; 00276 00277 // check stop criteria 00278 if (target) { 00279 if (nrValidClusters != opt->clusters) { 00280 break; 00281 } 00282 else { 00283 if (nrClusters - nrValidClusters > invalid) { 00284 break; 00285 } 00286 } 00287 } 00288 else { 00289 if (nrValidClusters == opt->clusters) { 00290 target = true; 00291 invalid = nrClusters - nrValidClusters; 00292 } 00293 } 00294 00295 } 00296 00297 cout << "==================================" << endl; 00298 cout << "Clustering process finished" << endl; 00299 cout << "==================================" << endl; 00300 if (opt->runs > 1) { 00301 cout << "Check out the last run and second last run." << endl; 00302 cout << "Kernel Width Found: " << prev_q << endl; 00303 } 00304 cout << "Soft Constraint Estimate: " << prev_C << endl; 00305 cout << "Overall time elapsed: " << overall_time << " secs" << endl; 00306 // cout << "Memory used: " << sizeof(svc) + sizeof(kwg) << endl; 00307 cout << "Original command line: "; 00308 printUsedCommandLine(argc, argv); 00309 00310 delete svc; 00311 delete kwg; 00312 free(opt); 00313 }
Here is the call graph for this function:
| struct svc_options * parseCommandLineArgs | ( | int | argc, | |
| char ** | argv | |||
| ) | [read] |
Parse the command line arguments and return the options for the SVC running.
| argc | The number of arguments | |
| argv | The arguments |
Definition at line 446 of file svc.c.
References svc_options::alg_type, svc_options::bsv, BSV_CLASSIC, svc_options::C, svc_options::clusters, DEFAULT_ALG, DEFAULT_C, DEFAULT_EDIST, DEFAULT_KERNEL, DEFAULT_KW, DEFAULT_L_OFFSET, DEFAULT_RUNS, DEFAULT_SOFTENING, DEFAULT_VALID_THRESHOLD, svc_options::detfile, svc_options::e_dist, svc_options::filename, svc_options::kernel, svc_options::l_offset, svc_options::q, svc_options::runs, svc_options::softening, usage(), and svc_options::valid_threshold.
Referenced by main().
00446 { 00447 00448 using namespace damina; 00449 00450 struct svc_options *opt; 00451 opt = (struct svc_options *)malloc(sizeof(struct svc_options)); 00452 00453 opt->filename = (char *)malloc(2 * sizeof(char)); 00454 strcpy(opt->filename, ""); 00455 00456 opt->q = DEFAULT_KW; 00457 opt->C = DEFAULT_C; 00458 opt->alg_type = DEFAULT_ALG; 00459 opt->runs = DEFAULT_RUNS; 00460 opt->l_offset = DEFAULT_L_OFFSET; 00461 opt->valid_threshold = DEFAULT_VALID_THRESHOLD; 00462 opt->softening = DEFAULT_SOFTENING; 00463 opt->detfile = NULL; 00464 opt->e_dist = DEFAULT_EDIST; 00465 opt->kernel = DEFAULT_KERNEL; 00466 opt->bsv = BSV_CLASSIC; 00467 00468 bool file_f = true; 00469 bool cluster_f = false; 00470 00471 int ch; 00472 00473 /* options descriptor */ 00474 static struct option longopts[] = { 00475 { "kernel-width", required_argument, NULL, 'q' }, 00476 { "soft-margin", required_argument, NULL, 'C' }, 00477 { "num-clusters", required_argument, NULL, 'c' }, 00478 { "labeling-alg", required_argument, NULL, 'l' }, 00479 { "dataset-file", required_argument, NULL, 'f' }, 00480 { "runs", required_argument, NULL, 'r' }, 00481 { "labels-offset", required_argument, NULL, 'o' }, 00482 { "valid-threshold",required_argument, NULL, 't' }, 00483 { "softening", required_argument, NULL, 's' }, 00484 { "dump-details", required_argument, NULL, 'D' }, 00485 { "euclidean-distance", required_argument, NULL, 'e' }, 00486 { "kernel-type", required_argument, NULL, 'k' }, 00487 { "bsv-clustering", required_argument, NULL, 'b' } 00488 00489 }; 00490 00491 while ((ch = getopt_long(argc, argv, "q:C:c:l:f:r:o:t:s:D:e:k:b:", longopts, NULL)) != -1) { 00492 switch (ch) { 00493 00494 case 'b': { 00495 opt->bsv = (int)strtol(optarg, (char**)NULL, 10); 00496 //if ((errno == EINVAL) || (opt->bsv != (int)BSV_CLASSIC) || (opt->bsv != (int)BSV_ENHANCED)) 00497 // err(1, "%s is an invalid value for BSV Clustering Policy. It must be an inter value in [%d,%d].", optarg, BSV_CLASSIC, BSV_ENHANCED); 00498 break; 00499 } 00500 00501 00502 case 'k': { 00503 opt->kernel = (int)strtol(optarg, (char**)NULL, 10); 00504 //if ((errno == EINVAL) || (opt->kernel != KernelType::gaussian) || (opt->kernel != KernelType::laplacian)) 00505 // err(1, "%s is an invalid value for Kernel Type. It must be %d or %d.", optarg, KernelType::gaussian, KernelType::laplacian); 00506 break; 00507 } 00508 00509 case 'e': { 00510 opt->e_dist = (int)strtol(optarg, (char**)NULL, 10); 00511 if ((errno == EINVAL) || (opt->e_dist < 1) || (opt->e_dist > 2)) 00512 err(1, "%s is an invalid value for Euclidean Distance type. It must be an inter value in [1,2].", optarg); 00513 break; 00514 } 00515 00516 case 'q': { 00517 opt->q = (double)strtod(optarg, (char**)NULL); 00518 if ((errno == EINVAL) || (opt->q <= 0)) 00519 err(1, "%s is an invalid value for Kernel Width. It must be a real value greater than 0.", optarg); 00520 break; 00521 } 00522 00523 case 'C': { 00524 opt->C = (double)strtod(optarg, (char**)NULL); 00525 if ((errno == EINVAL) || (opt->C <= 0)) 00526 err(1, "%s is an invalid value for Soft Margin Constant. It must be a real value greater than 0.", optarg); 00527 break; 00528 } 00529 00530 case 'c': { 00531 opt->clusters = (int)strtol(optarg, (char**)NULL, 10); 00532 if ((errno == EINVAL) || (opt->clusters <= 0)) 00533 err(1, "%s is an invalid value for Number of Clusters. It must be an integer value greater than 0.", optarg); 00534 00535 cluster_f = true; 00536 break; 00537 } 00538 00539 case 'r': { 00540 opt->runs = (int)strtol(optarg, (char**)NULL, 10); 00541 if ((errno == EINVAL) || (opt->runs <= 0)) 00542 err(1, "%s is an invalid value for Number of Runs. It must be an integer value greater than 0.", optarg); 00543 break; 00544 } 00545 00546 case 's': { 00547 opt->softening = (double)strtod(optarg, (char**)NULL); 00548 if ((errno == EINVAL) || (opt->softening <= 0)) 00549 err(1, "%s is an invalid value for the Softening Constant. It must be an real value greater than 0.", optarg); 00550 break; 00551 } 00552 00553 case 't': { 00554 opt->valid_threshold = (int)strtol(optarg, (char**)NULL, 10); 00555 if ((errno == EINVAL) || (opt->valid_threshold <= 0)) 00556 err(1, "%s is an invalid value for Minimum Clusters Cardinality. It must be an integer value greater than 0.", optarg); 00557 break; 00558 } 00559 00560 case 'o': { 00561 opt->l_offset = (int)strtol(optarg, (char**)NULL, 10); 00562 if ((errno == EINVAL) || (opt->l_offset < 0)) 00563 err(1, "%s is an invalid value for Labels Offset. It must be an integer value greater than or equal to 0.", optarg); 00564 break; 00565 } 00566 00567 case 'l': { 00568 opt->alg_type = (int)strtol(optarg, (char**)NULL, 10); 00569 if ((errno == EINVAL) || (opt->alg_type > 1) || (opt->alg_type < 0)) 00570 err(1, "%s is an invalid value for Cluster Labeling algorithm. It must be 0 or 1.", optarg); 00571 break; 00572 } 00573 00574 case 'f': { 00575 if ((open(optarg, O_RDONLY, 0)) == -1) { 00576 err(1, "Unable to open %s", optarg); 00577 break; 00578 } 00579 00580 free(opt->filename); 00581 opt->filename = (char *)malloc((strlen(optarg) + 1) * sizeof(char)); 00582 strcpy(opt->filename, optarg); 00583 00584 file_f = true; 00585 break; 00586 } 00587 00588 case 'D': { 00589 free(opt->detfile); 00590 opt->detfile = (char *)malloc((strlen(optarg) + 1) * sizeof(char)); 00591 strcpy(opt->detfile, optarg); 00592 break; 00593 } 00594 00595 default: { 00596 usage(argv); 00597 } 00598 } 00599 } 00600 argc -= optind; 00601 argv += optind; 00602 00603 if ((file_f == false) || (cluster_f == false)) { 00604 return NULL; 00605 } 00606 return opt; 00607 }
Here is the call graph for this function:
Here is the caller graph for this function:
| void printAllPoints | ( | damina::SVClustering * | clust, | |
| unsigned long int | classOffset, | |||
| char * | filename | |||
| ) |
Print all points with their original classes and cluster assigned.
| clust | The involved Support Vector Clustering engine | |
| classOffset | The label of the first class in the original classification |
Definition at line 322 of file svc.c.
References damina::SVClustering::getClustersAssignment(), and damina::SVClustering::getOriginalClasses().
Referenced by main().
00322 { 00323 00324 filebuf fb; 00325 fb.open(filename, ios::app); 00326 ostream out(&fb); 00327 00328 std::vector<int> labels = clust->getClustersAssignment(); // clustering result 00329 std::vector<int> orig_c = clust->getOriginalClasses((int)classOffset); // original classification 00330 00331 for (unsigned long int i = 0; i < labels.size(); i++) { 00332 out << "Point #" << i << " - " << "Cluster: " << labels[i] << ", Class: " << orig_c[i] << endl; 00333 } 00334 out << endl; 00335 fb.close(); 00336 }
Here is the call graph for this function:
Here is the caller graph for this function:
| void printUsedCommandLine | ( | int | argc, | |
| char * | argv[] | |||
| ) |
Print the whole command line used to invoke the program.
| opt | The result of the command line parsing |
Definition at line 646 of file svc.c.
Referenced by main().
00646 { 00647 cout << argv[0] << " "; 00648 00649 for (int i = 1; i <= argc; i++) { 00650 cout << argv[i] << " "; 00651 } 00652 cout << endl << endl; 00653 }
Here is the caller graph for this function:
| void usage | ( | char ** | argv | ) |
Print the usage on the standard error.
| argv | The argv parameter passed on the comand line |
Definition at line 614 of file svc.c.
References BSV_CLASSIC, BSV_ENHANCED, CCL, CG, DEFAULT_ALG, DEFAULT_C, DEFAULT_L_OFFSET, DEFAULT_RUNS, DEFAULT_SOFTENING, DEFAULT_VALID_THRESHOLD, L1, and L2.
Referenced by main(), and parseCommandLineArgs().
00614 { 00615 using namespace damina; 00616 00617 cerr << endl << "Usage: " << argv[0] << " -q (0,+inf) -C (0,+inf) -b [" << BSV_CLASSIC << "," << BSV_ENHANCED << "] -k" << KernelType::gaussian << "," << KernelType::laplacian << " -e [1,2] -c [1, N] -l [0-1] -f <dataset-filename> -D <detail-filename> -r [1, N] -o [0, N] -t [1,N] -s (0,1]" << endl << endl; 00618 cerr << "\t -q: the kernel width value must be a real value greater than zero - DEFAULT: auto" << endl; 00619 cerr << "\t -C: the soft margin constant must be a real value greater than zero - DEFAULT: " << DEFAULT_C << endl; 00620 cerr << "\t -b: the BSV Clustering Policy - DEFAULT: " << BSV_ENHANCED << endl; 00621 cerr << "\t\t " << BSV_CLASSIC << " for classic policy" << endl; 00622 cerr << "\t\t " << BSV_ENHANCED << " for enhanced policy (aim to get better results)" << endl; 00623 cerr << "\t -k: the Kernel Type - DEFAULT: " << KernelType::gaussian << endl; 00624 cerr << "\t\t " << KernelType::gaussian << " for Gaussian Kernel (K(x,y) = exp(-gamma * ||x-y||^2))" << endl; 00625 cerr << "\t\t " << KernelType::laplacian << " for Laplacian Kernel (K(x,y) = exp(-gamma*|x-y|_1)) " << endl; 00626 cerr << "\t -e: the Euclidean Measure to use in vector-distance measurement - DEFAULT: " << L2 << endl; 00627 cerr << "\t\t " << L1 << " for L1 distance (sum |x_i - y_i|) " << endl; 00628 cerr << "\t\t " << L2 << " for L2 distance (sum sqrt(||x_i - y_i||^2)) " << endl; 00629 cerr << "\t -l: the cluster labeling algorithm type - DEFAULT: " << DEFAULT_ALG << endl; 00630 cerr << "\t\t " << CCL << " for Cone Cluster Labeling" << endl; 00631 cerr << "\t\t " << CG << " for Complete Graph Cluster Labeling" << endl; 00632 cerr << "\t -c: the number of cluster to request must be an integer value greater than zero" << endl; 00633 cerr << "\t -f: the filename of the dataset file, in LIBSVM format" << endl; 00634 cerr << "\t -D: the filename of the file to store detailed clusters assignment" << endl; 00635 cerr << "\t -r: the number of runs to perform - DEFAULT: " << DEFAULT_RUNS << endl; 00636 cerr << "\t -o: the labels offset (i.e. the label of the first class) - DEFAULT: " << DEFAULT_L_OFFSET << endl; 00637 cerr << "\t -s: the softening constant for kernel width generation - DEFAULT: " << DEFAULT_SOFTENING << endl; 00638 cerr << "\t -t: the minimum clusters cardinality - DEFAULT: " << DEFAULT_VALID_THRESHOLD << endl << endl; 00639 }
Here is the caller graph for this function:
1.5.2