Data Mining. Concepts and Techniques, 3rd Edition

HAN

22-ind-673-708-9780123814791

2011/6/1

3:27

Page 680

#8

680

Index

complex data types (Continued)

summary, 586

symbolic sequence data, 586, 588–590

time-series data, 586, 587–588

composite join indices, 162

compressed patterns, 281

mining, 307–312

mining by pattern clustering, 308–310

compression, 100, 120

lossless, 100

lossy, 100

theory, 601

computer science applications, 613

concept characterization, 180

concept comparison, 180

concept description, 166, 180

concept hierarchies, 142, 179

for generalizing data, 150

illustrated, 143, 144

implicit, 143

manual provision, 144

multilevel association rule mining with, 285

multiple, 144

for nominal attributes, 284

for specializing data, 150

concept hierarchy generation, 112, 113, 120

based on number of distinct values, 118

illustrated, 112

methods, 117–119

for nominal data, 117–119

with prespecified semantic connections, 119

schema, 119

conditional probability table (CPT), 394, 395–396

confidence, 21

association rule, 21

interval, 219–220

limits, 373

rule, 245, 246

conflict resolution strategy, 356

confusion matrix, 365–366, 386

illustrated, 366

connectionist learning, 398

consecutive rules, 92

Constrained Vector Quantization Error (CVQE)

algorithm, 536

constraint-based clustering, 447, 497, 532–538, 539

categorization of constraints and, 533–535

hard constraints, 535–536

methods, 535–538

soft constraints, 536–537

speeding up, 537–538

See also cluster analysis

constraint-based mining, 294–301, 320

interactive exploratory mining/analysis, 295

as mining trend, 623

constraint-based patterns/rules, 281

constraint-based sequential pattern mining, 589

constraint-guided mining, 30

constraints

antimonotonic, 298, 301

association rule, 296–297

cannot-link, 533

on clusters, 533

coherence, 535

conflicting, 535

convertible, 299–300

data, 294

data-antimonotonic, 300

data-pruning, 300–301, 320

data-succinct, 300

dimension/level, 294, 297

hard, 534, 535–536, 539

inconvertible, 300

on instances, 533, 539

interestingness, 294, 297

knowledge type, 294

monotonic, 298

must-link, 533, 536

pattern-pruning, 297–300, 320

rules for, 294

on similarity measures, 533–534

soft, 534, 536–537, 539

succinct, 298–299

content-based retrieval, 596

context indicators, 314

context modeling, 316

context units, 314

contextual attributes, 546, 573

contextual outlier detection, 546–547, 582

with identified context, 574

normal behavior modeling, 574–575

structures as contexts, 575

summary, 575

transformation to conventional outlier

detection, 573–574

contextual outliers, 545–547, 573, 581

example, 546, 573

mining, 573–575

contingency tables, 95

continuous attributes, 44

contrasting classes, 15, 180

initial working relations, 177

prime relation, 175, 177

convertible constraints, 299–300

---

HAN

22-ind-673-708-9780123814791

2011/6/1

3:27

Page 681

#9

Index

681

COP k-means algorithm, 536

core descendants, 305

colossal patterns, 306

merging of core patterns, 306

core patterns, 304–305

core ratio, 305

correlation analysis, 94

discretization by, 117

interestingness measures, 264

with lift, 266–267

nominal data, 95–96

numeric data, 96–97

redundancy and, 94–98

correlation coefficient, 94, 96

numeric data, 96–97

correlation rules, 265, 272

correlation-based clustering methods, 511

correlations, 18

cosine measure, 268

cosine similarity, 77

between two term-frequency vectors, 78

cost complexity pruning algorithm, 345

cotraining, 432–433

covariance, 94, 97

numeric data, 97–98

CPAR. See Classification based on Predictive

Association Rules

credit policy analysis, 608–609

CRM. See customer relationship management

crossover operation, 426

cross-validation, 370–371, 386

k-fold, 370

leave-one-out, 371

in number of clusters determination, 487

stratified, 371

cube gradient analysis, 321

cube shells, 192, 211

computing, 211

cube space

discovery-driven exploration, 231–234

multidimensional data analysis in, 227–234

prediction mining in, 227

subspaces, 228–229

cuboid trees, 205

cuboids, 137

apex, 111, 138, 158

base, 111, 137–138, 158

child, 193

individual, 190

lattice of, 139, 156, 179, 188–189,

234, 290

sparse, 190

subset selection, 160

See also data cubes

curse of dimensionality, 158, 179

customer relationship management (CRM),

619

customer retention analysis, 610

CVQE. See Constrained Vector Quantization Error

algorithm

cyber-physical systems (CPS), 596, 623–624

D

data

antimonotonicity, 300

archeology, 6

biological sequence, 586, 590–591

complexity, 32

conversion to knowledge, 2

cyber-physical system, 596

for data mining, 8

data warehouse, 13–15

database, 9–10

discrimination, 16

dredging, 6

generalizing, 150

graph, 14

growth, 2

linearly inseparable, 413–415

linearly separated, 409

multimedia, 14, 596

multiple sources, 15, 32

multivariate, 556

networked, 14

overfitting, 330

relational, 10

sample, 219

similarity and dissimilarity measures, 65–78

skewed, 47, 271

spatial, 14, 595

spatiotemporal, 595–596

specializing, 150

statistical descriptions, 44–56

streams, 598

symbolic sequence, 586, 588–589

temporal, 14

text, 14, 596–597

time-series, 586, 587

“tombs,” 5

training, 18

transactional, 13–14

types of, 33

web, 597–598

data auditing tools, 92

---