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 694 #22 694 Index neighborhoods density, 471 distance-based outlier detection, 560 k-distance, 565 nested loop algorithm, 561, 562 networked data, 14 networks, 592 heterogeneous, 592, 593 homogeneous, 592, 593 information, 592–594 mining in science applications, 612–613 social, 592 statistical modeling of, 592–594 neural networks, 19, 398 backpropagation, 398–408 as black boxes, 406 for classification, 19, 398 disadvantages, 406 fully connected, 399, 406–407 learning, 398 multilayer feed-forward, 398–399 pruning, 406–407 rule extraction algorithms, 406, 407 sensitivity analysis, 408 three-layer, 399 topology definition, 400 two-layer, 399 neurodes, 399 Ng-Jordan-Weiss algorithm, 521, 522 no materialization, 159 noise filtering, 318 noisy data, 89–91 nominal attributes, 41 concept hierarchies for, 284 correlation analysis, 95–96 dissimilarity between, 69 example, 41 proximity measures, 68–70 similarity computation, 70 values of, 79, 288
nonlinear SVMs, 413–415 nonparametric statistical methods, 553–558
nonvolatile data warehouses, 127 normalization, 112, 120 data transformation by, 113–115 by decimal scaling, 115 min-max, 114
null rules, 92 null-invariant measures, 270–271, 272 null-transactions, 270, 272 number of, 270 problem, 292–293 numeric attributes, 43–44, 79 covariance analysis, 98 interval-scaled, 43, 79 ratio-scaled, 43–44, 79 numeric data, dissimilarity on, 72–74 numeric prediction, 328, 385 classification, 328 support vector machines (SVMs) for, 408 numerosity reduction, 86, 100, 120 techniques, 100 O object matching, 94 objective interestingness measures, 21–22 one-class model, 571–572 one-pass cube computation, 198 one-versus-all (OVA), 430 online analytical mining (OLAM), 155, 227 online analytical processing (OLAP), 4, 33, 128, 179 access patterns, 129 data contents, 128 database design, 129 dice operation, 148 drill-across operation, 148 drill-down operation, 11, 135–136, 146 drill-through operation, 148 example operations, 147 functionalities of, 154 hybrid OLAP, 164–165, 179 indexing, 125, 160–163 in information networks, 594 in knowledge discovery process, 125 market orientation, 128 multidimensional (MOLAP), 132, 164, 179 OLTP versus, 128–129, 130 operation integration, 125 operations, 146–148 pivot (rotate) operation, 148 queries, 129, 130, 163–164 query processing, 125, 163–164 relational OLAP, 132, 164, 165, 179 roll-up operation, 11, 135–136, 146 sample data effectiveness, 219 server architectures, 164–165 servers, 132 slice operation, 148 spatial, 595 statistical databases versus, 148–149 HAN 22-ind-673-708-9780123814791 2011/6/1 3:27 Page 695 #23 Index 695 user-control versus automation, 167 view, 129 online transaction processing (OLTP), 128 access patterns, 129 customer orientation, 128 data contents, 128 database design, 129 OLAP versus, 128–129, 130 view, 129 operational metadata, 135 OPTICS, 473–476 cluster ordering, 474–475, 477 core-distance, 475 density estimation, 477 reachability-distance, 475 structure, 476 terminology, 476 See also cluster analysis; density-based methods ordered attributes, 103 ordering class-based, 358 dimensions, 210 rule, 357 ordinal attributes, 42, 79 dissimilarity between, 75 example, 42 proximity measures, 74–75 outlier analysis, 20–21 clustering-based techniques, 66 example, 21 in noisy data, 90 spatial, 595 outlier detection, 543–584 angle-based (ABOD), 580 application-specific, 548–549 categories of, 581 CELL method, 562–563 challenges, 548–549 clustering analysis and, 543 clustering for, 445 clustering-based methods, 552–553, 560–567 collective, 548, 575–576 contextual, 546–547, 573–575 distance-based, 561–562 extending, 577–578 global, 545 handling noise in, 549 in high-dimensional data, 576–580, 582 with histograms, 558–560 intrusion detection, 569–570 methods, 549–553 mixture of parametric distributions, 556–558 multivariate, 556 novelty detection relationship, 545 proximity-based methods, 552, 560–567, 581 semi-supervised methods, 551 statistical methods, 552, 553–560, 581 supervised methods, 549–550 understandability, 549 univariate, 554 unsupervised methods, 550 outlier subgraphs, 576 outliers
angle-based, 20, 543, 544, 580 collective, 547–548, 581 contextual, 545–547, 573, 581 density-based, 564 distance-based, 561 example, 544 global, 545, 581 high-dimensional, modeling, 579–580 identifying, 49 interpretation of, 577 local proximity-based, 564–565 modeling, 548 in small clusters, 571 types of, 545–548, 581 visualization with boxplot, 555 oversampling, 384, 386 example, 384–385
pairwise alignment, 590 pairwise comparison, 372 PAM. See Partitioning Around Medoids algorithm parallel and distributed data-intensive mining algorithms, 31 parallel coordinates, 59, 62 parametric data reduction, 105–106 parametric statistical methods, 553–558 Pareto distribution, 592 partial distance method, 425 partial materialization, 159–160, 179, 234 strategies, 192 partition matrix, 538 partitioning algorithms, 451–457 in Apriori efficiency, 255–256 bootstrapping, 371, 386 criteria, 447 cross-validation, 370–371, 386 Gini index and, 342 holdout method, 370, 386 random sampling, 370, 386
Yüklə 7,95 Mb. Dostları ilə paylaş:
|