Data Mining. Concepts and Techniques, 3rd Edition
HAN 22-ind-673-708-9780123814791
Yüklə 7,95 Mb. Pdf görüntüsü
|
HAN 22-ind-673-708-9780123814791 2011/6/1 3:27 Page 674 #2 674 Index applications (Continued) targeted, 27–28 telecommunications industry, 611 Web search engines, 28 application-specific outlier detection, 548–549 approximate patterns, 281 mining, 307–312 Apriori algorithm, 248–253, 272 dynamic itemset counting, 256 efficiency, improving, 254–256 example, 250–252 hash-based technique, 255 join step, 249 partitioning, 255–256 prune step, 249–250 pseudocde, 253 sampling, 256 transaction reduction, 255 Apriori property, 194, 201, 249 antimonotonicity, 249 in Apriori algorithm, 298 Apriori pruning method, 194 arrays
3-D for dimensions, 196 sparse compression, 198–199 association analysis, 17–18 association rules, 245 approximate, 281 Boolean, 281 compressed, 281 confidence, 21, 245, 246, 416 constraint-based, 281 constraints, 296–297 correlation, 265, 272 discarded, 17 fittest, 426 frequent patterns and, 280 generation from frequent itemsets, 253, 254 hybrid-dimensional, 288 interdimensional, 288 intradimensional, 287 metarule-guided mining of, 295–296 minimum confidence threshold, 18, 245 minimum support threshold, 245 mining, 272 multidimensional, 17, 287–289, 320 multilevel, 281, 283–287, 320 near-match, 281 objective measures, 21 offspring, 426 quantitative, 281, 289, 320 redundancy-aware top-k, 281 single-dimensional, 17, 287 spatial, 595 strong, 264–265, 272 support, 21, 245, 246, 417 top-k, 281 types of values in, 281 associative classification, 415, 416–419, 437 CBA, 417 CMAR, 417–418 CPAR, 418–419 rule confidence, 416 rule support, 417 steps, 417 asymmetric binary dissimilarity, 71 asymmetric binary similarity, 71 attribute construction, 112 accuracy and, 105 multivariate splits, 344 attribute selection measures, 331, 336–344 CHAID, 343 gain ratio, 340–341 Gini index, 341–343 information gain, 336–340 Minimum Description Length (MDL), 343–344
multivariate splits, 343–344 attribute subset selection, 100, 103–105 decision tree induction, 105 forward selection/backward elimination combination, 105 greedy methods, 104–105 stepwise backward elimination, 105 stepwise forward selection, 105 attribute vectors, 40, 328 attribute-oriented induction (AOI), 166–178, 180 algorithm, 173 for class comparisons, 175–178 for data characterization, 167–172 data generalization by, 166–178 generalized relation, 172 implementation of, 172–174 attributes, 9, 40 abstraction level differences, 99 behavioral, 546, 573 binary, 41–42, 79 Boolean, 41 categorical, 41 class label, 328 contextual, 546, 573 continuous, 44 correlated, 54–56 dimension correspondence, 10
HAN 22-ind-673-708-9780123814791 2011/6/1 3:27 Page 675 #3 Index 675 discrete, 44 generalization, 169–170 generalization control, 170 generalization threshold control, 170 grouping, 231 interval-scaled, 43, 79 of mixed type, 75–77 nominal, 41, 79 numeric, 43–44, 79 ordered, 103 ordinal, 41, 79 qualitative, 41 ratio-scaled, 43–44, 79 reducts of, 427 removal, 169 repetition, 346 set of, 118 splitting, 333 terminology for, 40 type determination, 41 types of, 39 unordered, 103 audio data mining, 604–607, 624 automatic classification, 445 AVA. See all-versus-all AVC-group, 347 AVC-set, 347 average() , 215
B background knowledge, 30–31 backpropagation, 393, 398–408, 437 activation function, 402 algorithm illustration, 401 biases, 402, 404 case updating, 404 efficiency, 404 epoch updating, 404 error, 403 functioning of, 400–403 hidden layers, 399 input layers, 399 input propagation, 401–402 interpretability and, 406–408 learning, 400 learning rate, 403–404 logistic function, 402 multilayer feed-forward neural network, 398–399
network pruning, 406–407 neural network topology definition, 400 output layers, 399 sample learning calculations, 404–406 sensitivity analysis, 408 sigmoid function, 402 squashing function, 403 terminating conditions, 404 unknown tuple classification, 406 weights initialization, 401 See also classification bagging, 379–380 algorithm illustration, 380 boosting versus, 381–382 in building random forests, 383 bar charts, 54 base cells, 189 base cuboids, 111, 137–138, 158 Basic Local Alignment Search Tool (BLAST), 591 Baum-Welch algorithm, 591 Bayes’ theorem, 350–351 Bayesian belief networks, 393–397, 436 algorithms, 396 components of, 394 conditional probability table (CPT), 394, 395
directed acyclic graph, 394–395 gradient descent strategy, 396–397 illustrated, 394 mechanisms, 394–396 problem modeling, 395–396 topology, 396 training, 396–397
Bayesian classification basis, 350 Bayes’ theorem, 350–351 class conditional independence, 350 naive, 351–355, 385 posterior probability, 351 prior probability, 351 BCubed precision metric, 488, 489 BCubed recall metric, 489 behavioral attributes, 546, 573 believability, data, 85 BI (business intelligence), 27 biases, 402, 404 biclustering, 512–519, 538 application examples, 512–515 enumeration methods, 517, 518–519 gene expression example, 513–514 methods, 517–518 optimization-based methods, 517–518 recommender system example, 514–515 types of, 538 Yüklə 7,95 Mb. Dostları ilə paylaş:
|