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 688 #16 688 Index frequent patterns (Continued) closed, 262–264, 280 concepts, 243–244 constraint-based, 281 dimensions, 281 diversity, 280 exploration, 313–319 growth, 257–259, 272 max, 262–264, 280 mining, 243–244, 279–325 mining constraints or criteria, 281 number of dimensions involved in, 281 semantic annotation of, 313–317 sequential, 243 strong associations, 437 structured, 243 trees, 257–259 types of values in, 281 frequent subgraphs, 591 front-end client layer, 132 full materialization, 159, 179, 234 fuzzy clustering, 499–501, 538 data set for, 506 with EM algorithm, 505–507 example, 500 expectation step (E-step), 505 flexibility, 501 maximization step (M-step), 506–507 partition matrix, 499 as soft clusters, 501 fuzzy logic, 428 fuzzy sets, 428–429, 437, 499 evaluation, 500–501 example, 499
gain ratio, 340 C4.5 use of, 340 formula, 341 maximum, 341 gateways, 131 gene expression, 513–514 generalization attribute, 169–170 attribute, control, 170 attribute, threshold control, 170 in multimedia data mining, 596 process, 172 results presentation, 174 synchronous, 175 generalized linear models, 599–600 generalized relations attribute-oriented induction, 172 presentation of, 174 threshold control, 170 generative model, 467–469 genetic algorithms, 426–427, 437 genomes, 15 geodesic distance, 525–526, 539 diameter, 525 eccentricity, 525 measurements based on, 526 peripheral vertex, 525 radius, 525 geographic data warehouses, 595 geometric projection visualization, 58–60 Gini index, 341 binary enforcement, 332 binary indexes, 341 CART use of, 341 decision tree induction using, 342–343 minimum, 342 partitioning and, 342 global constants, for missing values, 88 global outliers, 545, 581 detection, 545 example, 545 Google
Flu Trends, 2 popularity of, 619–620 gradient descent strategy, 396–397 algorithms, 397 greedy hill-climbing, 397 as iterative, 396–397 graph and network data clustering, 497, 522–532, 539 applications, 523–525 bipartite graph, 523 challenges, 523–525, 530 cuts and clusters, 529–530 generic method, 530–531 geodesic distance, 525–526 methods, 528–532 similarity measures, 525–528 SimRank, 526–528 social network, 524–525 web search engines, 523–524
graph cuts, 539 graph data, 14 graph index structures, 591 graph pattern mining, 591–592, 612–613 graphic displays data presentation software, 44–45 histogram, 54, 55 HAN 22-ind-673-708-9780123814791 2011/6/1 3:27 Page 689 #17 Index 689 quantile plot, 51–52 quantile-quantile plot, 52–54 scatter plot, 54–56 greedy hill-climbing, 397 greedy methods, attribute subset selection, 104–105 grid-based methods, 450, 479–483, 491 CLIQUE, 481–483 STING, 479–481 See also cluster analysis grid-based outlier detection, 562–564 CELL method, 562, 563 cell properties, 562 cell pruning rules, 563
group-based support, 286 group-by clause, 231 grouping attributes, 231 grouping variables, 231 Grubb’s test, 555
hamming distance, 431 hard constraints, 534, 539 example, 534 handling, 535–536 harmonic mean, 369 hash-based technique, 255 heterogeneous networks, 592 classification of, 593 clustering of, 593 ranking of, 593 heterogeneous transfer learning, 436 hidden Markov model (HMM), 590, 591 hierarchical methods, 449, 457–470, 491 agglomerative, 459–461 algorithmic, 459, 461–462 Bayesian, 459 BIRCH, 458, 462–466 Chameleon, 458, 466–467 complete linkages, 462, 463 distance measures, 461–462 divisive, 459–461 drawbacks, 449 merge or split points and, 458 probabilistic, 459, 467–470 single linkages, 462, 463 See also cluster analysis hierarchical visualization, 63 treemaps, 63, 65 Worlds-with-Worlds, 63, 64 high-dimensional data, 301 clustering, 447 data distribution of, 560 frequent pattern mining, 301–307 outlier detection in, 576–580, 582 row enumeration, 302 high-dimensional data clustering, 497, 508–522, 538, 553
biclustering, 512–519 dimensionality reduction methods, 510, 519–522 example, 508–509 problems, challenges, and methodologies, 508–510
subspace clustering methods, 509, 510–511
See also cluster analysis HilOut algorithm, 577–578 histograms, 54, 106–108, 116 analysis by discretization, 115–116 attributes, 106 binning, 106 construction, 559 equal-frequency, 107 equal-width, 107 example, 54 illustrated, 55, 107 multidimensional, 108 as nonparametric model, 559 outlier detection using, 558–560 holdout method, 370, 386 holistic measures, 145 homogeneous networks, 592 classification of, 593 clustering of, 593 Hopkins statistic, 484–485 horizontal data format, 259 hybrid OLAP (HOLAP), 164–165, 179 hybrid-dimensional association rules, 288
I IBM Intelligent Miner, 603, 606 iceberg condition, 191 iceberg cubes, 160, 179, 190, 235 BUC construction, 201 computation, 160, 193–194, 319 computation and storage, 210–211 computation with Star-Cubing algorithm, 204–210 materialization, 319 specification of, 190–191 See also data cubes icon-based visualization, 60 Chernoff faces, 60–61
Yüklə 7,95 Mb. Dostları ilə paylaş:
|