<<

. 133
( 137 .)



>>


communication channels, 89
Inmon), 474

customer relationships, 457
Business Modeling and Data Mining
efforts, 90
(Dorian Pyle), 60

good prospects, identifying, 88“89
Data Preparation for Data Mining
index-based scores, 92“95
(Dorian Pyle), 75

marketing campaigns
The Data Warehouse Toolkit (Ralph

acquisition-time variables, 110
Kimball), 474

credit risks, reducing exposure to,
Genetic Algorithms in Search,

113“114
Optimization, and Machine

cross-selling, 115“116
Learning (Goldberg), 445

customer response, tracking, 109
purchases, market based analysis, 289

customer segmentation, 111“113
purchasing frequencies, behavior-

differential response analysis,
based variables, 575“576

107“108
purity measures, splitting criteria,

discussed, 95
decision trees, 177“178

fixed budgets, 97“100
p-values, statistics, 126

new customer information,
Pyle, Dorian

gathering, 109“110 Business Modeling and Data Mining, 60

people most influenced by, 106“107 Data Preparation for Data Mining, 75

Index 637


relational database management
Q
system (RDBMS)

quadratic discriminates, box
discussed, 474

diagrams, 200

source systems, 594“595

quality of data, association rules, 308

star schema, 505

question asking, data exploration,

suppliers, 13

67“68

support, 511

Quinlan, J. Ross (Iterative

relevance feedback, MBR, 267“268

Dichotomiser 3), 190

replicating results, 33

q-values, statistics, 126

reporting requirements, OLAP,

R 495“496
resources
range values, statistics, 137

geographical, 555“556
rate plans, wireless telephone

optimization, generic algorithms,
services, 7

433“435
ratios

response

data transformation, 75

biased sampling, 146

lift ratio, 81“84

communication channels, 89

RDBMS. See relational database
control groups

management system
market research versus, 38

real estate appraisals, neural network
marketing campaigns, 106

example, 213“217
cumulative response

recall measurements, classification
concentration, 82“83

codes, 273“274
results, assessing, 85

recency, frequency, and monetary
customer relationships, 457

(RFM) value, 575
differential response analysis,

recommendation-based businesses,
marketing campaigns, 107“108

16“17
erroneous conclusions, 74

records

free text, 285

combining values within, 569

good response scores, 34

default classes, 194

marketing campaigns, 96“97

transactional, 574

prediction, MBR, 258

rectangular regions, decision trees, 197

proof-of-concept projects, 599

recursive algorithms, 173

response models

reduction in variance, splits, decision

generic algorithms, 440“443

trees, 183

prospects, ranking, 36

regression

response times, interactive

building models, 8

systems, 33

estimation tasks, 10

sample sizes, 145

linear, 139

single response rates, 141

regression trees, 170

survey response

statistics, 139

customer classification, 91

techniques, generic algorithms, 423

inconclusive, 46

638 Index


response, survey response (continued)
data quality, 308

profiling, 53
dissociation rules, 317

survey-based market research, 113
effectiveness of, 299“301

useful data sources, 61
inexplicable rules, 297“298

results
point-of-sale data, 288

actionable, 22
practical limits, overcoming,

assessing, 85
311“313

comparing expectations to, 31
prediction, 70

deliverables, data transformation,
probabilities, calculating, 309

57“58
products, hierarchical categories, 305

measuring, virtuous cycle, 30“32
sequential analysis, 318“319

neural networks, 241“243
for store comparisons, 315“316

replicating, 33
trivial rules, 297

statistical analysis, 141“143
virtual items, 307

tainted, 72
decision trees, 193“194

retention
generalized delta, 229

calculating, 385“386
rule-oriented problems, 176

churn and, 116“120

S
customer relationships, 467“469

SAC (Simplifying Assumptions
exponential decay, 389“390, 393

Corporation), 97, 100

hazards, 404“405

<<

. 133
( 137 .)



>>