Maximize profit and optimize decisions with advanced business analytics Profit-Driven Business Analytics provides actionable guidance on optimizing the use of data to add value and drive better business. Combining theoretical and technical insights into daily operations and long-term strategy, this book acts as a development manual for practitioners seeking to conceive, develop, and manage advanced analytical models. Detailed discussion delves into the wide range of analytical approaches and modeling techniques that can help maximize business payoff, and the author team draws upon their recent research to share deep insight about optimal strategy. Real-life case studies and examples illustrate these techniques at work, and provide clear guidance for implementation in your own organization. From step-by-step instruction on data handling, to analytical fine-tuning, to evaluating results, this guide provides invaluable guidance for practitioners seeking to reap the advantages of true business analytics.
Despite widespread discussion surrounding the value of data in decision making, few businesses have adopted advanced analytic techniques in any meaningful way. This book shows you how to delve deeper into the data and discover what it can do for your business.
- Reinforce basic analytics to maximize profits
- Adopt the tools and techniques of successful integration
- Implement more advanced analytics with a value-centric approach
- Fine-tune analytical information to optimize business decisions
Both data stored and streamed has been increasing at an exponential rate, and failing to use it to the fullest advantage equates to leaving money on the table. From bolstering current efforts to implementing a full-scale analytics initiative, the vast majority of businesses will see greater profit by applying advanced methods. Profit-Driven Business Analytics provides a practical guidebook and reference for adopting real business analytics techniques.
Table of Contents:
Foreword xv
Acknowledgments xvii
Chapter 1 A Value-Centric Perspective Towards Analytics 1
Introduction 1
Business Analytics 3
Profit-Driven Business Analytics 9
Analytics Process Model 14
Analytical Model Evaluation 17
Analytics Team 19
Profiles 19
Data Scientists 20
Conclusion 23
Review Questions 24
Multiple Choice Questions 24
Open Questions 25
References 25
Chapter 2 Analytical Techniques 28
Introduction 28
Data Preprocessing 29
Denormalizing Data for Analysis 29
Sampling 30
Exploratory Analysis 31
Missing Values 31
Outlier Detection and Handling 32
Principal Component Analysis 33
Types of Analytics 37
Predictive Analytics 37
Introduction 37
Linear Regression 38
Logistic Regression 39
Decision Trees 45
Neural Networks 52
Ensemble Methods 56
Bagging 57
Boosting 57
Random Forests 58
Evaluating Ensemble Methods 59
Evaluating Predictive Models 59
Splitting Up the Dataset 59
Performance Measures for Classification Models 63
Performance Measures for Regression Models 67
Other Performance Measures for Predictive Analytical
Models 68
Descriptive Analytics 69
Introduction 69
Association Rules 69
Sequence Rules 72
Clustering 74
Survival Analysis 81
Introduction 81
Survival Analysis Measurements 83
Kaplan Meier Analysis 85
Parametric Survival Analysis 87
Proportional Hazards Regression 90
Extensions of Survival Analysis Models 92
Evaluating Survival Analysis Models 93
Social Network Analytics 93
Introduction 93
Social Network Definitions 94
Social Network Metrics 95
Social Network Learning 97
Relational Neighbor Classifier 98
Probabilistic Relational Neighbor Classifier 99
Relational Logistic Regression 100
Collective Inferencing 102
Conclusion 102
Review Questions 103
Multiple Choice Questions 103
Open Questions 108
Notes 110
References 110
Chapter 3 Business Applications 114
Introduction 114
Marketing Analytics 114
Introduction 114
RFM Analysis 115
Response Modeling 116
Churn Prediction 118
X-selling 120
Customer Segmentation 121
Customer Lifetime Value 123
Customer Journey 129
Recommender Systems 131
Fraud Analytics 134
Credit Risk Analytics 139
HR Analytics 141
Conclusion 146
Review Questions 146
Multiple Choice Questions 146
Open Questions 150
Note 151
References 151
Chapter 4 Uplift Modeling 154
Introduction 154
The Case for Uplift Modeling: Response Modeling 155
Effects of a Treatment 158
Experimental Design, Data Collection, and Data
Preprocessing 161
Experimental Design 161
Campaign Measurement of Model Effectiveness 164
Uplift Modeling Methods 170
Two-Model Approach 172
Regression-Based Approaches 174
Tree-Based Approaches 183
Ensembles 193
Continuous or Ordered Outcomes 198
Evaluation of Uplift Models 199
Visual Evaluation Approaches 200
Performance Metrics 207
Practical Guidelines 210
Two-Step Approach for Developing Uplift Models 210
Implementations and Software 212
Conclusion 213
Review Questions 214
Multiple Choice Questions 214
Open Questions 216
Note 217
References 217
Chapter 5 Profit-Driven Analytical Techniques 220
Introduction 220
Profit-Driven Predictive Analytics 221
The Case for Profit-Driven Predictive Analytics 221
Cost Matrix 222
Cost-Sensitive Decision Making with Cost-Insensitive
Classification Models 228
Cost-Sensitive Classification Framework 231
Cost-Sensitive Classification 234
Pre-Training Methods 235
During-Training Methods 247
Post-Training Methods 253
Evaluation of Cost-Sensitive Classification Models 255
Imbalanced Class Distribution 256
Implementations 259
Cost-Sensitive Regression 259
The Case for Profit-Driven Regression 259
Cost-Sensitive Learning for Regression 260
During Training Methods 260
Post-Training Methods 261
Profit-Driven Descriptive Analytics 267
Profit-Driven Segmentation 267
Profit-Driven Association Rules 280
Conclusion 283
Review Questions 284
Multiple Choice Questions 284
Open Questions 289
Notes 290
References 291
Chapter 6 Profit-Driven Model Evaluation
and Implementation 296
Introduction 296
Profit-Driven Evaluation of Classification Models 298
Average Misclassification Cost 298
Cutoff Point Tuning 303
ROC Curve-Based Measures 310
Profit-Driven Evaluation with Observation-Dependent
Costs 334
Profit-Driven Evaluation of Regression Models 338
Loss Functions and Error-Based Evaluation Measures 339
REC Curve and Surface 341
Conclusion 345
Review Questions 347
Multiple Choice Questions 347
Open Questions 350
Notes 351
References 352
Chapter 7 Economic Impact 355
Introduction 355
Economic Value of Big Data and Analytics 355
Total Cost of Ownership (TCO) 355
Return on Investment (ROI) 357
Profit-Driven Business Analytics 359
Key Economic Considerations 359
In-Sourcing versus Outsourcing 359
On Premise versus the Cloud 361
Open-Source versus Commercial Software 362
Improving the ROI of Big Data and Analytics 364
New Sources of Data 364
Data Quality 367
Management Support 369
Organizational Aspects 370
Cross-Fertilization 371
Conclusion 372
Review Questions 373
Multiple Choice Questions 373
Open Questions 376
Notes 377
References 377
About the Authors 378
Index 381
About the Author :
WOUTER VERBEKE is assistant professor of Business Informatics and Data Analytics at Vrije Universiteit Brussel (Belgium). He is the coauthor of Fraud Analytics using Descriptive, Predictive, and Social Network Techniques.
BART BAESENS is a professor at KU Leuven (Belgium) and a lecturer at the University of Southampton (United Kingdom). He is the author of Credit Risk Management and Analytics in a Big Data World, as well as coauthor of Fraud Analytics using Descriptive, Predictive, and Social Network Techniques.
CRISTIÁN BRAVO is a lecturer vin business analytics in the department of Decision Analytics and Risk at the University of Southampton.
