proof-of-concept projects

expectations, 599

identifying, 599-601

implementation, 601-605 propensity

categorical variables, 242

propensity-to-respond score, 97 proportion

converting counts to, 75-76

difference of proportion chi-square tests versus, 153-154 statistical analysis, 143-144

penetration, 203

standard error of, statistical analysis, 139-141 proportional hazards

Cox, 410-111

discussed, 408

examples of, 409

limitations of, 411-412 proportional scoring, census data,

94-95 prospecting

advertising techniques, 90-94

communication channels, 89

customer relationships, 457

efforts, 90

good prospects, identifying, 88-89

index-based scores, 92-95

marketing campaigns acquisition-time variables, 110 credit risks, reducing exposure to,

113-114 cross-selling, 115-116 customer response, tracking, 109 customer segmentation, 111-113 differential response analysis,

107-108 discussed, 95 fixed budgets, 97-100 new customer information,

gathering, 109-110 people most influenced by, 106-107

planning, 27

profitability, 100-104

response modeling, 96-97

types of, 111

up-selling, 115-116 messages, selecting appropriate,

89-90 ranking, 88-89 roles in, 88 targeting, 88

time dependency and, 160 prospective customer value, 115 prototypes, proof-of-concept

projects, 599 pruning, decision trees

C5 algorithm, 190-191

CART algorithm, 185, 188-189

discussed, 184

minimum support pruning, 312 stability-based, 191-192 public records, house-hold level

data, 96 publications Building the Data Warehouse (Bill

Inmon), 474 Business Modeling and Data Mining

(Dorian Pyle), 60 Data Preparation for Data Mining

(Dorian Pyle), 75 The Data Warehouse Toolkit (Ralph

Kimball), 474 Genetic Algorithms in Search, Optimization, and Machine Learning (Goldberg), 445 purchases, market based analysis, 289 purchasing frequencies, behavior-

based variables, 575-576 purity measures, splitting criteria,

decision trees, 177-178 p-values, statistics, 126 Pyle, Dorian Business Modeling and Data Mining, 60 Data Preparation for Data Mining, 75

quadratic discriminates, box

diagrams, 200 quality of data, association rules, 308 question asking, data exploration,

67-68

Quinlan, J. Ross (Iterative

Dichotomiser 3), 190 q-values, statistics, 126

range values, statistics, 137 rate plans, wireless telephone

services, 7 ratios

data transformation, 75 lift ratio, 81-84 RDBMS. See relational database

management system real estate appraisals, neural network

example, 213-217 recall measurements, classification

codes, 273-274 recency, frequency, and monetary

(RFM) value, 575 recommendation-based businesses,

16-17 records

combining values within, 569

default classes, 194

transactional, 574 rectangular regions, decision trees, 197 recursive algorithms, 173 reduction in variance, splits, decision

trees, 183 regression

building models, 8

estimation tasks, 10

linear, 139

regression trees, 170 statistics, 139

techniques, generic algorithms, 423

relational database management system (RDBMS) discussed, 474 source systems, 594-595 star schema, 505 suppliers, 13 support, 511 relevance feedback, MBR, 267-268 replicating results, 33 reporting requirements, OLAP,

495-496 resources geographical, 555-556 optimization, generic algorithms, 433-435 response biased sampling, 146 communication channels, 89 control groups market research versus, 38 marketing campaigns, 106 cumulative response concentration, 82-83 results, assessing, 85 customer relationships, 457 differential response analysis,

marketing campaigns, 107-108 erroneous conclusions, 74 free text, 285 good response scores, 34 marketing campaigns, 96-97 prediction, MBR, 258 proof-of-concept projects, 599 response models generic algorithms, 440-443 prospects, ranking, 36 response times, interactive

systems, 33 sample sizes, 145 single response rates, 141 survey response customer classification, 91 inconclusive, 46

response, survey response (continued) profiling, 53

survey-based market research, 113

useful data sources, 61 results actionable, 22 assessing, 85

comparing expectations to, 31 deliverables, data transformation, 57-58

measuring, virtuous cycle, 30-32 neural networks, 241-243 replicating, 33 statistical analysis, 141-143 tainted, 72 retention calculating, 385-386 churn and, 116-120 customer relationships, 467-169 exponential decay, 389-390, 393 hazards, 404-405

median customer lifetime value, 387 retention curves, 386-389 truncated mean lifetime value, 389 retrospective customer value, 115 revenue, behavior-based variables,

581-585 revolvers, behavior-based

variables, 580 RFM (recency, frequency, and

monetary) value, 575 ring diagrams, as alternative to

decision trees, 199-201 risks hazards, 403

proof-of-concept projects, 599 ROC curves, 98-99, 101 root sets, link analysis, 333 RuleQuest Web site, 190 rules association rules actionable rules, 296 affinity grouping, 11 anonymous versus identified transactions, 308

data quality, 308 dissociation rules, 317 effectiveness of, 299-301 inexplicable rules, 297-298 point-of-sale data, 288 practical limits, overcoming,

311-313 prediction, 70

probabilities, calculating, 309 products, hierarchical categories, 305 sequential analysis, 318-319 for store comparisons, 315-316 trivial rules, 297 virtual items, 307 decision trees, 193-194 generalized delta, 229 rule-oriented problems, 176

SAC (Simplifying Assumptions

Corporation), 97, 100 sample sizes, statistical analysis, 145 sample variation, statistics, 129 SAS Enterprise Miner Tree Viewer

tool, 167-168 scalability, data mining, 533-534 scaling, automatic cluster detection,

363-364 scanners, point-of-sale, 3 scarce data, 62

SCF (sectional center facility), 553 schemata, generic algorithms, 434,

436-438 scores

bizocity, 112-113

cutoff, 98

decision trees, 169-170 good response, 34 index-based, 92-95 model deployment, 84-85 propensity-to-respond, 97 proportional, census data, 94-95 score sets, 52

scoring platforms, data mining, 527-528