Statistical Modeling and Analysis for Database Marketing: Effective Techniques for Mining Big Data

by Ratner; Bruce

Edition: 1st

ISBN13: 9781574443448

ISBN10: 1574443445

Format: Hardcover

Pub. Date: 2003-05-28

Publisher(s): Chapman & Hall/

Other versions by this Author

List Price: ~~$99.95~~

Rent Textbook

Select for Price

Add to Cart

There was a problem. Please try again later.

New Textbook

We're Sorry
Sold Out

Used Textbook

We're Sorry
Sold Out

eTextbook

We're Sorry
Not Available

Buy from our Marketplace starting at $115.83

Summary

Traditional statistical methods are limited in their ability to meet the modern challenge of mining large amounts of data. Data miners, analysts, and statisticians are searching for innovative new data mining techniques with greater predictive power, an attribute critical for reliable models and analyses.Statistical Modeling and Analysis for Database Marketing: Effective Techniques for Mining Big Data delivers a collection of successful database marketing methodologies for big data. This compendium solves common database marketing problems by applying new hybrid modeling techniques that combine traditional statistical and new machine learning methods. The book delivers a thorough analysis of these cutting-edge techniques, which include non-statistical machine learning and genetic intelligent hybrid models.By following the step-by-step procedures detailed in the text, database marketing professionals can learn how to apply the proper statistical techniques to any database marketing challenge. The practical case studies and examples provided involve real problems and real data, and are taken from a variety of industries, including banking, insurance, finance, retail, and telecommunications.

Introduction
The Personal Computer and Statistics	p. 1
Statistics and Data Analysis	p. 3
EDA	p. 4
The EDA Paradigm	p. 6
EDA Weaknesses	p. 7
Small and Big Data	p. 8
Data Size Characteristics	p. 8
Data Size: Personal Observation of One	p. 9
Data Mining Paradigm	p. 9
Statistics and Machine Learning	p. 11
Statistical Learning	p. 12
Two Simple Data Mining Methods for Variable Assessment
Correlation Coefficient	p. 15
Scatterplots	p. 17
Data Mining	p. 18
Example #1	p. 18
Example #2	p. 21
Smoothed Scatterplot	p. 21
General Association Test	p. 26
Summary	p. 27
Logistic Regression: The Workhorse of Database Response Modeling
Logistic Regression Model	p. 32
Illustration	p. 32
Scoring a LRM	p. 33
Case Study	p. 35
Candidate Predictor and Dependent Variables	p. 36
Logits and Logit Plots	p. 36
Logits for Case Study	p. 37
The Importance of Straight Data	p. 38
Re-expressing for Straight Data	p. 39
Ladder of Powers	p. 39
Bulging Rule	p. 40
Measuring Straight Data	p. 41
Straight Data for Case Study	p. 42
Re-expressing FD2_OPEN	p. 44
Re-expressing INVESTMENT	p. 44
Techniques When Bulging Rule Does Not Apply	p. 46
Fitted Logit Plot	p. 46
Smooth Predicted vs. Actual Plot	p. 47
Re-expressing MOS_OPEN	p. 47
Smooth Predicted vs. Actual Plot for MOS_OPEN	p. 49
Assessing the Importance of Variables	p. 50
Computing the G Statistic	p. 52
Importance of a Single Variable	p. 53
Importance of a Subset of Variables	p. 53
Comparing the Importance of Different Subsets of Variables	p. 53
Important Variables for Case Study	p. 54
Importance of the Predictor Variables	p. 55
Relative Importance of the Variables	p. 56
Selecting the Best Subset	p. 57
Best Subset of Variables for Case Study	p. 58
Visual Indicators of Goodness of Model Predictions	p. 59
Smooth Residual by Score Groups Plot	p. 59
Smooth Residual by Score Groups Plot for Case Study	p. 60
Smooth Actual vs. Predicted by Decile Groups Plot	p. 62
Smooth Actual vs. Predicted by Decile Groups Plot for Case Study	p. 63
Smooth Actual vs. Predicted by Score Groups Plot	p. 65
Smooth Actual vs. Predicted by Score Groups Plot for Case Study	p. 65
Evaluating the Data Mining Work	p. 68
Comparison of Smooth Residual by Score Groups Plots: EDA vs. NonEDA Models	p. 69
Comparison of Smooth Actual vs. Predicted by Decile Groups Plots: EDA vs. NonEDA Models	p. 71
Comparison of Smooth Actual vs. Predicted by Score Groups Plots: EDA vs. NonEDA Models	p. 71
Summary of the Data Mining Work	p. 71
Smoothing a Categorical Variable	p. 74
Smoothing FD_TYPE with CHAID	p. 75
Importance of CH_FTY_1 and CH_FTY_2	p. 78
Additional Data Mining Work for Case Study	p. 78
Comparison of Smooth Residual by Score Group Plots: 4var- vs. 3var-EDA Models	p. 79
Comparison of Smooth Actual vs. Predicted by Decile Groups Plots: 4var- vs. 3var-EDA Models	p. 81
Comparison of Smooth Actual vs. Predicted by Score Groups Plots: 4var- vs. 3var-EDA Models	p. 82
Final Summary of the Additional Data Mining Work	p. 84
Summary	p. 85
Ordinary Regression: The Workhorse of Database Profit Modeling
Ordinary Regression Model	p. 87
Illustration	p. 88
Scoring A OLS Profit Model	p. 89
Mini Case Study	p. 91
Straight Data for Mini Case Study	p. 91
Re-expressing INCOME	p. 93
Re-expressing AGE	p. 95
Smooth Predicted vs. Actual Plot	p. 96
Assessing the Importance of Variables	p. 98
Defining the F Statistic and R-squared	p. 98
Importance of a Single Variable	p. 99
Importance of a Subset of Variables	p. 99
Comparing the Importance of Different Subsets of Variables	p. 99
Important Variables for Mini Case Study	p. 100
Relative Importance of the Variables	p. 101
Selecting the Best Subset	p. 101
Best Subset of Variable for Case Study	p. 102
PROFIT Model with gINCOME and AGE	p. 103
Best PROFIT Model	p. 106
Suppressor Variable AGE	p. 106
Summary	p. 108
CHAID for Interpreting a Logistic Regression Model
Logistic Regression Model	p. 111
Database Marketing Response Model Case Study	p. 112
Odds Ratio	p. 113
CHAID	p. 114
Proposed CHAID-Based Method	p. 114
Multivariable CHAID Trees	p. 117
CHAID Market Segmentation	p. 121
CHAID Tree Graphs	p. 123
Summary	p. 126
The Importance of the Regression Coefficient
The Ordinary Regression Model	p. 129
Four Questions	p. 130
Important Predictor Variables	p. 130
P-Values and Big Data	p. 132
Returning to Question #1	p. 132
Predictor Variable's Effect on Prediction	p. 133
The Caveat	p. 134
Returning to Question #2	p. 136
Ranking Predictor Variables by Effect On Prediction	p. 136
Returning to Question #3	p. 138
Returning to Question #4	p. 138
Summary	p. 139
The Predictive Contribution Coefficient: A Measure of Predictive Importance
Background	p. 141
Illustration of Decision Rule	p. 143
Predictive Contribution Coefficient	p. 145
Calculation of Predictive Contribution Coefficient	p. 146
Extra Illustration of Predictive Contribution Coefficient	p. 148
Summary	p. 152
CHAID for Specifying a Model with Interaction Variables
Interaction Variables	p. 155
Strategy for Modeling with Interaction Variables	p. 156
Strategy Based on the Notion of a Special Point	p. 156
Example of a Response Model with an Interaction Variable	p. 157
CHAID for Uncovering Relationships	p. 159
Illustration of CHAID for Specifying a Model	p. 160
An Exploratory Look	p. 164
Database Implication	p. 165
Summary	p. 166
Market Segment Classification Modeling with Logistic Regression
Binary Logistic Regression	p. 169
Necessary Notation	p. 170
Polychotomous Logistic Regression Model	p. 171
Model Building with PLR	p. 172
Market Segmentation Classification Model	p. 172
Survey of Cellular Phone Users	p. 173
CHAID Analysis	p. 174
CHAID Tree Graphs	p. 177
Market Segment Classification Model	p. 180
Summary	p. 182
CHAID as a Method for Filling in Missing Values
Introduction to the Problem of Missing Data	p. 185
Missing-Data Assumption	p. 188
CHAID Imputation	p. 189
Illustration	p. 190
CHAID Mean-Value Imputation for a Continuous Variable	p. 191
Many Mean-Value CHAID Imputations for a Continuous Variable	p. 192
Regression-Tree Imputation for LIF_DOL	p. 193
CHAID Most-Likely Category Imputation for a Categorical Variable	p. 196
CHAID Most-Likely Category Imputation for GENDER	p. 196
Classification Tree Imputation for GENDER	p. 198
Summary	p. 200
Identifying Your Best Customers: Descriptive, Predictive and Look-Alike Profiling
Some Definitions	p. 203
Illustration of a Flawed Targeting Effort	p. 204
Well-Defined Targeting Effort	p. 205
Predictive Profiles	p. 208
Continuous Trees	p. 212
Look-Alike Profiling	p. 215
Look-Alike Tree Characteristics	p. 216
Summary	p. 219
Assessment of Database Marketing Models
Accuracy for Response Model	p. 221
Accuracy for Profit Model	p. 222
Decile Analysis and Cum Lift for Response Model	p. 223
Decile Analysis and Cum Lift for Profit Model	p. 227
Precision for Response Model	p. 229
Precision for Profit Model	p. 231
Construction of SWMAD	p. 231
Separability for Response and Profit Models	p. 233
Guidelines for Using Cum Lift, HL/SWMAD and CV	p. 233
Summary	p. 234
Bootstrapping in Database Marketing: A New Approach for Validating Models
Traditional Model Validation	p. 237
Illustration	p. 238
Three Questions	p. 239
The Bootstrap	p. 240
Traditional Construction of Confidence Intervals	p. 241
How to Bootstrap	p. 242
Simple Illustration	p. 242
Bootstrap Decile Analysis Validation	p. 244
Another Question	p. 245
Bootstrap Assessment of Model Implementation Performance	p. 246
Illustration	p. 249
Bootstrap Assessment of Model Efficiency	p. 253
Summary	p. 255
Visualization of Database Models
Brief History of the Graph	p. 257
Star Graph Basics	p. 258
Illustration	p. 260
Star Graphs for Single Variables	p. 261
Star Graphs for Many Variables Considered Jointly	p. 262
Profile Curves Method	p. 264
Profile Curves Basics	p. 264
Profile Analysis	p. 265
Illustration	p. 265
Profile Curves for RESPONSE Model	p. 269
Decile-Group Profile Curves	p. 271
Summary	p. 274
SAS Code for Star Graphs for Each Demographic Variable about the Deciles	p. 275
SAS Code for Star Graphs for Each Decile about the Demographic Variables	p. 277
SAS Code for Profile Curves: All Deciles	p. 281
Genetic Modeling in Database Marketing: The GenIQ Model
What Is Optimization?	p. 285
What Is Genetic Modeling?	p. 286
Genetic Modeling: An Illustration	p. 287
Reproduction	p. 290
Crossover	p. 290
Mutation	p. 292
Parameters for Controlling a Genetic Model Run	p. 293
Genetic Modeling: Strengths and Limitations	p. 293
Goals of Modeling in Database Marketing	p. 294
The GenIQ Response Model	p. 295
The GenIQ Profit Model	p. 295
Case Study--Response Model	p. 296
Case Study--Profit Model	p. 299
Summary	p. 302
Finding the Best Variables for Database Marketing Models
Background	p. 303
Weakness in the Variable Selection Methods	p. 305
Goals of Modeling in Database Marketing	p. 307
Variable Selection with GenIQ	p. 308
GenIQ Modeling	p. 310
GenIQ-Structure Identification	p. 312
GenIQ Variable Selection	p. 316
Nonlinear Alternative to Logistic Regression Model	p. 316
Summary	p. 321
Interpretation of Coefficient-Free Models
The Linear Regression Coefficient	p. 323
Illustration for the Simple Ordinary Regression Model	p. 324
Illustration for the Simple Logistic Regression Model	p. 325
The Quasi-Regression Coefficient for Simple Regression Models	p. 326
Illustration of Quasi-RC for the Simple Ordinary Regression Model	p. 326
Illustration of Quasi-RC for the Simple Logistic Regression Model	p. 328
Illustration of Quasi-RC for Nonlinear Predictions	p. 331
Partial Quasi-RC for the Everymodel	p. 331
Calculating the Partial Quasi-RC for the Everymodel	p. 333
Illustration for the Multiple Logistic Regression Model	p. 335
Quasi-RC for a Coefficient-Free Model	p. 340
Illustration of Quasi-RC for a Coefficient-Free Model	p. 341
Summary	p. 348
Index	p. 351
Table of Contents provided by Rittenhouse. All Rights Reserved.

Kids

Men

Unisex

Women

For You

For Your Car

For Your Home

For Your Pet

For Your Tech

Artwork

Games

Holiday

Mascot

Office Decor

Spirit

Stationary

Diploma Frames

Graduation Gifts

Office Supplies

Clothing

Accessories

Gifts

Statistical Modeling and Analysis for Database Marketing: Effective Techniques for Mining Big Data

Rent Textbook

New Textbook

Used Textbook

eTextbook

Summary

Table of Contents

Statistical Modeling and Analysis for Database Marketing: Effective Techniques for Mining Big Data

Rent Textbook

New Textbook

Used Textbook

eTextbook

How Marketplace Works:

Summary

Table of Contents

Digital License