Decision Tree
뿌리튼튼 CS/Data Mining2015. 4. 23. 16:52
난이도 ★★★★☆ (쉽진 않지만 apriori 보다는 쉽다)
목표
information gain을 사용하여 decision tree를 만들고, 그것을 이용하여 test set을 분류하라.
요구사항
● 실행파일 이름 : dt.exe
● 실행프로그램 인자 : training file name, test file name
● 훈련파일 형식
● 테스트파일 형식
● 기타 주의사항
1. Attribute 선택은 information gain값으로 한다.
2. 정확도가 높을 수록 높은 점수를 받는다.
● 결과 테스트용 입력 파일
1. Training file
dt_train.txt2. Test file
● 결과 테스트용 출력 파일
● 소스 코드
1. 직접 다운
dt.cpp2. 소스 보기
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 | #include <stdio.h> #include <math.h> // log2 #include <vector> #include <iostream> #include <fstream> #include <string> #include <sstream> #pragma warning(disable:4996) #define DEBUG_MODE false // if true, print logs using namespace std; typedef struct { vector<string> attributeNames; vector<vector<string>> records; int classLabelIndex; } Table; typedef struct Node { int attributeIndex; vector<struct Child> children; } Node; typedef struct Child { string value; // attribute value struct Node* pNode; string classValue; // only for leaves } Child; typedef struct { string classValue; int count; } ClassValueCountPair; // only for building tree typedef struct { vector<vector<string>*> remainingRecords; vector<ClassValueCountPair> classValueCountPairs; // for majarity voting. If size() == 1, it means that all records for the remaining records belong to the same class. } ChildData; // only for getting gains typedef struct { string attributeValue; vector<ClassValueCountPair> classValueCountPairs; } AttributeData; // I(pi[]) double getInfo(const vector<int>* pi) { double sum = 0.0; for (int i = 0; i < pi->size(); i++) { sum += pi->at(i); } double ret = 0.0; for (int i = 0; i < pi->size(); i++) { if (pi->at(i) == 0) { continue; } ret += (-1 * pi->at(i) / sum) * log2(pi->at(i) / sum); } return ret; } // Gain(attribute) double getGain(const int attributeIndex, const int classLableIndex, const vector<vector<string>*>* remainingRecords) { vector<ClassValueCountPair> classValueTotalCountPairs; // for infoD vector<AttributeData> attributeData; // for infoAD for (int i = 0; i < remainingRecords->size(); i++) { const string classValue = remainingRecords->at(i)->at(classLableIndex); const string attributeValue = remainingRecords->at(i)->at(attributeIndex); // make attributeData bool attributeValueFound = false; for (int j = 0; j < attributeData.size(); j++) { if (attributeData[j].attributeValue.compare(attributeValue) == 0) { bool found = false; for (int k = 0; k < attributeData[j].classValueCountPairs.size(); k++) { if (attributeData[j].classValueCountPairs[k].classValue.compare(classValue) == 0) { attributeData[j].classValueCountPairs[k].count++; found = true; break; } } if (!found) { ClassValueCountPair classValueCountPair; classValueCountPair.classValue = classValue; classValueCountPair.count = 1; attributeData[j].classValueCountPairs.push_back(classValueCountPair); } attributeValueFound = true; break; } } if (!attributeValueFound) { AttributeData attributeDatum; attributeDatum.attributeValue = attributeValue; ClassValueCountPair classValueCountPair; classValueCountPair.classValue = classValue; classValueCountPair.count = 1; attributeDatum.classValueCountPairs.push_back(classValueCountPair); attributeData.push_back(attributeDatum); } // make classValueTotalCountPairs bool found = false; for (int j = 0; j < classValueTotalCountPairs.size(); j++) { if (classValueTotalCountPairs[j].classValue.compare(classValue) == 0) { classValueTotalCountPairs[j].count++; found = true; break; } } if (!found) { ClassValueCountPair classValueCountPair; classValueCountPair.classValue = classValue; classValueCountPair.count = 1; classValueTotalCountPairs.push_back(classValueCountPair); } } // get infoD double infoD; { vector<int> pi; for (int i = 0; i < classValueTotalCountPairs.size(); i++) { pi.push_back(classValueTotalCountPairs[i].count); } infoD = getInfo(&pi); } // get infoAD's sum double infoADSum = 0.0; for (int i = 0; i < attributeData.size(); i++) { vector<int> pi; double sum = 0.0; for (int j = 0; j < attributeData[i].classValueCountPairs.size(); j++) { pi.push_back(attributeData[i].classValueCountPairs[j].count); sum += attributeData[i].classValueCountPairs[j].count; } infoADSum += (sum / remainingRecords->size()) * getInfo(&pi); } if (DEBUG_MODE) { printf("[rec_size=%2d, att_id=%d, att_size=%d] infoD=%.3f, infoADSum=%.3f, gain=%.3f\n", remainingRecords->size(), attributeIndex, attributeData.size(), infoD, infoADSum, infoD - infoADSum); } return infoD - infoADSum; } // get decision tree by recursion Node* getTree(const vector<vector<string>*> remainingRecords, vector<int> remainingAttributeIndexes, const int classLableIndex) { // select an attribute which has max gain if (DEBUG_MODE) { printf("** getGain **\n"); } int selected = 0; double maxGain = getGain(remainingAttributeIndexes[selected], classLableIndex, &remainingRecords); for (int i = 1; i < remainingAttributeIndexes.size(); i++) { double gain = getGain(remainingAttributeIndexes[i], classLableIndex, &remainingRecords); if (gain > maxGain) { maxGain = gain; selected = i; } } const int selectedAttributeIndex = remainingAttributeIndexes[selected]; if (DEBUG_MODE) { printf("\n\n"); } // get new remainingAttributeIndexes which excludes the selectedAttributeIndex remainingAttributeIndexes.erase(remainingAttributeIndexes.begin() + selected); // make node (1) : node allocation and set node.attributeIndex Node* node = new Node; node->attributeIndex = selectedAttributeIndex; // make node (2) : make children and childData vector<ChildData> childData; for (int i = 0; i < remainingRecords.size(); i++) { string childValue = remainingRecords[i]->at(selectedAttributeIndex); string classValue = remainingRecords[i]->at(classLableIndex); bool childValueFound = false; for (int j = 0; j < node->children.size(); j++) { // found : already exists if (node->children[j].value.compare(childValue) == 0) { childData[j].remainingRecords.push_back(remainingRecords[i]); // renew classValueCountPairs bool classValueFound = false; for (int k = 0; k < childData[j].classValueCountPairs.size(); k++) { if (childData[j].classValueCountPairs[k].classValue.compare(classValue) == 0) { childData[j].classValueCountPairs[k].count++; classValueFound = true; break; } } if (!classValueFound) { ClassValueCountPair classValueCountPair; classValueCountPair.classValue = classValue; classValueCountPair.count = 1; childData[j].classValueCountPairs.push_back(classValueCountPair); } childValueFound = true; break; } } // not found : create new child and new childDatum if (!childValueFound) { // create new child which will be a part of tree Child child; child.value = childValue; node->children.push_back(child); // create new childDatum which will be used as a parameter when building tree ChildData childDatum; childDatum.remainingRecords.push_back(remainingRecords[i]); ClassValueCountPair classValueCountPair; classValueCountPair.classValue = classValue; classValueCountPair.count = 1; childDatum.classValueCountPairs.push_back(classValueCountPair); childData.push_back(childDatum); } } // make node (3) : get children's nodes by recursion for (int i = 0; i < node->children.size(); i++) { // if all records for a given node belong to the same class if (childData[i].classValueCountPairs.size() == 1) { node->children[i].classValue = childData[i].classValueCountPairs[0].classValue; node->children[i].pNode = NULL; } // if there are no remaining attributes for further partitioning, majority voting is employed for classifying the leaf else if (remainingAttributeIndexes.empty()) { int maxClassValueIndex = 0; for (int j = 1; j < childData[i].classValueCountPairs.size(); j++) { if (childData[i].classValueCountPairs[j].count > childData[i].classValueCountPairs[maxClassValueIndex].count) { maxClassValueIndex = j; } } node->children[i].classValue = childData[i].classValueCountPairs[maxClassValueIndex].classValue; node->children[i].pNode = NULL; } // if normal case else { node->children[i].pNode = getTree(childData[i].remainingRecords, remainingAttributeIndexes, classLableIndex); } } return node; } // get class value by recursive tree traversal void makeDecision(vector<string>* record, const Node* node) { const string attributeValue = record->at(node->attributeIndex); if (DEBUG_MODE) { cout << attributeValue << "\t"; } for (int i = 0; i < node->children.size(); i++) { if (node->children[i].value.compare(attributeValue) == 0) { if (node->children[i].pNode == NULL) { record->push_back(node->children[i].classValue); } else { makeDecision(record, node->children[i].pNode); } break; } } } int main(int argc, char** argv) { if (argc != 3) { cout << "Error Code: 001" << "\n" << "argc should be 3" << endl; return 0; } char* trainingFileName = argv[1]; char* testFileName = argv[2]; // get training table Table trainingTable; ifstream trainingFile(trainingFileName); if (trainingFile.is_open()) { string line; // get names of attributes if (getline(trainingFile, line)) { stringstream ss(line); string s; while (ss >> s) { trainingTable.attributeNames.push_back(s); } } // get records while (getline(trainingFile, line)) { stringstream ss(line); string s; vector<string> record; while (ss >> s) { record.push_back(s); } trainingTable.records.push_back(record); } trainingFile.close(); } else { cout << "Error Code: 002" << "\n" << "cannot open training file. (name: " << trainingFileName << ")" << endl; return 0; } // get Decision Tree (1) : fill remainingRecords with whole records vector<vector<string>*> remainingRecords; for (int i = 0; i < trainingTable.records.size(); i++) { remainingRecords.push_back(&trainingTable.records[i]); } // get Decision Tree (2) : fill remainingAttributeIndexes with whole indexes except class label index trainingTable.classLabelIndex = trainingTable.attributeNames.size() - 1; vector<int> remainingAttributeIndexes; for (int i = 0; i < trainingTable.attributeNames.size(); i++) { if (i != trainingTable.classLabelIndex) { remainingAttributeIndexes.push_back(i); } } // get Decision Tree (3) : get decision tree by calling a recursive function Node* root = getTree(remainingRecords, remainingAttributeIndexes, trainingTable.classLabelIndex); // get test sets vector<vector<string>> testRecords; ifstream testFile(testFileName); if (testFile.is_open()) { string line; // skip names of attributes if (getline(testFile, line)) { // do nothing } // get records while (getline(testFile, line)) { stringstream ss(line); string s; vector<string> record; while (ss >> s) { record.push_back(s); } // get class value and save at the last column if (DEBUG_MODE) { printf("** makeDecision **\n"); } makeDecision(&record, root); if (DEBUG_MODE) { printf("\n\n"); } // push testRecords.push_back(record); } testFile.close(); } else { cout << "Error Code: 003" << "\n" << "cannot open test file. (name: " << testFileName << ")" << endl; return 0; } // output const string OUTPUT_FILE_NAME = "dt_result.txt"; ofstream outputFile(OUTPUT_FILE_NAME); if (outputFile.is_open()) { // print attribute names for (int i = 0; i < trainingTable.attributeNames.size(); i++) { outputFile << trainingTable.attributeNames[i]; if (i < trainingTable.attributeNames.size() - 1) { outputFile << "\t"; } else { outputFile << endl; } } // print attribute values and class values for (int i = 0; i < testRecords.size(); i++) { for (int j = 0; j < testRecords[i].size(); j++) { outputFile << testRecords[i][j]; if (j < testRecords[i].size() - 1) { outputFile << "\t"; } else { outputFile << endl; } } } outputFile.close(); } else { cout << "Error Code: 004" << "\n" << "cannot open output file. (name: " << OUTPUT_FILE_NAME << ")" << endl; return 0; } return 0; } | cs |
● 출처 : 2015년 한양대학교 김상욱 교수님의 데이터마이닝 수업 과제
'뿌리튼튼 CS > Data Mining' 카테고리의 다른 글
| Apriori 알고리즘 (0) | 2015.04.10 |
|---|
Apriori 알고리즘
뿌리튼튼 CS/Data Mining2015. 4. 10. 14:27
난이도 ★★★★★ (너무 어렵다)
요구사항
● 실행파일 이름 : apriori.exe
● 실행프로그램 인자 : minimum support, input file name, output file name
● 입력파일 형식
● 출력파일 형식
● 기타 주의사항
1. 실행시 인자로 넘겨주는 Min Support값과 출력시 Support값은 %값이다.
2. 출력파일에서의 행별 정렬 순서는 중요하지 않다.
● 결과 테스트용 입력 파일
● 결과 테스트용 출력 파일
1. Min Support = 10%
2. Min Support = 9%
3. Min Support = 6%
4. Min Support = 2% (굉장히 오래걸림. 1시간 이상)
● 소스 코드 (답은 맞지만 느립니다. 최적의 코드가 결코 아닙니다)
1. 직접 다운
2. 소스 보기
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 | #include <stdio.h> #include <stdlib.h> #include <vector> #include <iostream> #include <fstream> #include <string> #include <sstream> #include <set> #include <algorithm> #include <iomanip> #define DEBUG_MODE true #pragma warning(disable:4996) using namespace std; typedef struct { set<int> itemset; int sup; } Itemset; void makeC(vector<Itemset> *pC, vector<Itemset> *pLi) { pC->clear(); // self-joining L for (int i = 0; i < pLi->size() - 1; i++) { for (int j = i + 1; j < pLi->size(); j++) { // get L(i) - L(j) set<int> i_diff_j; set_difference(pLi->at(i).itemset.begin(), pLi->at(i).itemset.end(), pLi->at(j).itemset.begin(), pLi->at(j).itemset.end(), inserter(i_diff_j, i_diff_j.end())); // if could if (i_diff_j.size() == 1) { // generate a candidate set<int> candidate; candidate.insert(pLi->at(j).itemset.begin(), pLi->at(j).itemset.end()); candidate.insert(i_diff_j.begin(), i_diff_j.end()); // redundancy check bool found = false; for each (Itemset c in *pC) { set<int> result; set_difference(c.itemset.begin(), c.itemset.end(), candidate.begin(), candidate.end(), inserter(result, result.end())); if (result.empty()) { found = true; break; } } // if not found, push if (!found) { Itemset item; item.itemset = candidate; item.sup = 0; pC->push_back(item); } } } } // if don't need to be pruned, return if ((pC->size() < 1) || (pC->at(0).itemset.size() < 3)) { return; } // pruning int i = 0; while (true) { if (i >= pC->size()) { break; } for (int j = 0; j < pC->at(i).itemset.size(); j++) { // get subSet which excludes jth number set<int> subSet; set<int>::iterator it = pC->at(i).itemset.begin(); if (j == 0) { // first std::advance(it, 1); subSet.insert(it, pC->at(i).itemset.end()); } else if (j == pC->at(i).itemset.size() - 1) { // last std::advance(it, j); subSet.insert(pC->at(i).itemset.begin(), it); } else { // else std::advance(it, j); subSet.insert(pC->at(i).itemset.begin(), it); std::advance(it, 1); subSet.insert(it, pC->at(i).itemset.end()); } // check bool found = false; for each (Itemset l in *pLi) { set<int> result; set_difference(subSet.begin(), subSet.end(), l.itemset.begin(), l.itemset.end(), inserter(result, result.end())); if (result.empty()) { found = true; break; } } if (!found) { pC->erase(pC->begin() + i); i--; break; } } i++; } } void makeL(vector<vector<Itemset>> *pL, vector<Itemset> *pC, int minSupport) { vector<Itemset> Li; for each (Itemset c in *pC) { if (c.sup >= minSupport) { Itemset item; item.itemset = c.itemset; // could generate error item.sup = c.sup; Li.push_back(item); } } pL->push_back(Li); if (DEBUG_MODE) { printf("\n************************\n"); int i = pL->size() - 1; printf("L[%d] (size = %d, minSup = %d)\n", i, pL->at(i).size(), minSupport); for each (Itemset l in pL->at(i)) { printf("{ "); for each (int num in l.itemset) { printf("%d ", num); } printf(" }, %d\t", l.sup); } printf("\n*************************\n"); } } set<set<int>> findSubsets(set<int> A) { if (A.empty()) { set<set<int>> ret; return ret; } set<int> tmp; set<int>::iterator it = A.begin(); std::advance(it, 1); tmp.insert(it, A.end()); set<set<int>> B = findSubsets(tmp); set<set<int>> C; int A0 = *A.begin(); for each (set<int> D in B) { set<int> temp; temp.insert(A0); temp.insert(D.begin(), D.end()); C.insert(temp); set<int> temp2; temp2.insert(A0); C.insert(temp2); } if (B.empty()) { set<int> temp; temp.insert(A0); C.insert(temp); } B.insert(C.begin(), C.end()); return B; } int getSupport(set<int> *itemset, vector<vector<Itemset>> *pL) { for each (Itemset l in pL->at(itemset->size() - 1)) { set<int> result; set_difference(l.itemset.begin(), l.itemset.end(), itemset->begin(), itemset->end(), inserter(result, result.end())); if (result.empty()) { return l.sup; } } return -1; } int main(int argc, char** argv) { if (argc != 4) { cout << "Error Code: 001" << "\n" << "argc should be 4" << endl; return 0; } int relativeSupport = atoi(argv[1]); char* inputFileName = argv[2]; char* outputFileName = argv[3]; if (DEBUG_MODE) { printf("%d %s %s\n", relativeSupport, argv[2], argv[3]); } int sizeOfTransactions = 0; // scan to increment the count of all candidates in C(1) vector<vector<Itemset>> L; // L(i) : frequent itemset of size i vector<Itemset> C; // C(i) : candidate itemset of size i ifstream inputFile(inputFileName); if (inputFile.is_open()) { string line; while (getline(inputFile, line)) { sizeOfTransactions++; // get transaction set<int> transaction; stringstream ss(line); int num; while (ss >> num) { transaction.insert(num); // find num in C int target = -1; for (int i = 0; i < C.size(); i++) { if (C[i].itemset.find(num) != C[i].itemset.end()) { target = i; break; } } // not found if (target == -1) { Itemset item; item.itemset.insert(num); item.sup = 0; C.push_back(item); } // found else { C[target].sup++; } } // get C's sup for each (Itemset c in C) { set<int> result; set_difference(c.itemset.begin(), c.itemset.end(), transaction.begin(), transaction.end(), inserter(result, result.end())); if (result.empty()) { c.sup++; } } } inputFile.close(); } else { cout << "Error Code: 002" << "\n" << "Cannot open input file. (name: " << inputFileName << ")" << endl; return 0; } // get minSupport int minSupport = ceil(((double)sizeOfTransactions / 100) * relativeSupport); // get L(1) makeL(&L, &C, minSupport); // get C(i) & L(i) where i > 1 int i = 0; while (true) { // get C(i+1) makeC(&C, &L[i]); // if cannot generate C(i+1), break if (C.empty()) { break; } // scan to increment the count of all candidates in C ifstream inputFile(inputFileName); if (inputFile.is_open()) { string line; while (getline(inputFile, line)) { // get transaction set<int> transaction; stringstream ss(line); int num; while (ss >> num) { transaction.insert(num); } // get C's sup for (int j = 0; j < C.size(); j++) { set<int> result; set_difference(C[j].itemset.begin(), C[j].itemset.end(), transaction.begin(), transaction.end(), inserter(result, result.end())); if (result.empty()) { C[j].sup++; } } } inputFile.close(); } else { cout << "Error Code: 003(" << i << ")" << "\n" << "Cannot open input file. (name: " << inputFileName << ")" << endl; return 0; } // get L(i+1) makeL(&L, &C, minSupport); i++; } // output ofstream outputFile(outputFileName); if (outputFile.is_open()) { printf("\n*** output... ***\n"); outputFile << fixed << setprecision(2); for (int i = 1; i < L.size(); i++) { for (int j = 0; j < L.at(i).size(); j++) { // L.at(i) : {itemset}, {itemset}, {itemset}, ... // get all sizes of subset set<set<int>> subsets = findSubsets(L.at(i).at(j).itemset); // L.at(i).at(j) : {itemset} for each (set<int> subset1 in subsets) { set<int> subset2; set_difference(L.at(i).at(j).itemset.begin(), L.at(i).at(j).itemset.end(), subset1.begin(), subset1.end(), inserter(subset2, subset2.end())); if (subset2.empty()) { continue; } // print subset1 outputFile << "{"; int pos = 0; for each (int item in subset1) { outputFile << item; if (pos < subset1.size() - 1) { outputFile << ","; } else { outputFile << "}\t"; } pos++; } // print subset2 outputFile << "{"; pos = 0; for each (int item in subset2) { outputFile << item; if (pos < subset2.size() - 1) { outputFile << ","; } else { outputFile << "}\t"; } pos++; } // print relative support of original set (subset1 + subset2) outputFile << (L.at(i).at(j).sup / (double)sizeOfTransactions) * 100 << "\t"; // print confidence (original set's support / subset1's support) outputFile << (L.at(i).at(j).sup / (double)getSupport(&subset1, &L)) * 100 << endl; } } } outputFile.close(); } else { cout << "Error Code: 004" << "\n" << "Cannot open output file. (name: " << outputFileName << ")" << endl; return 0; } return 0; } | cs |
● 출처 : 2015년 한양대학교 김상욱 교수님의 데이터마이닝 수업 과제
'뿌리튼튼 CS > Data Mining' 카테고리의 다른 글
| Decision Tree (0) | 2015.04.23 |
|---|
