电信行业数据挖掘红皮书(pdf 194页)

Contents
Preface   vii
The tEAM that wrote this redbook . . . . . . . . vii
Special notice  . . . . . . . . . . . . . . . . . . . . . . . . ix
IBM trademarks  . . . . . . . . . . . . . . . . . . . . . . ix
Comments welcome . . . . . . . . . . . . . . . . . . . ix
Chapter 1. Introduction . . . . . . . . . . . . . . . 1
1.1 Why you should mine your own business . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 What are the telecoms business issues to address?. . . . . . . . . . . . . . . 3
1.3 How this book is structured  . . . . . . 4
1.4 Who should read this book?  . . . . . . 6
Chapter 2. Business Intelligence archITecture overvIEw . . . . . . . . . . . . . . . 7
2.1 Business Intelligence  . . . . . . . . . . . 8
2.2 Data warehouse  . . . . . . . . . . . . . . . 8
2.2.1 Data sources  . . . . . . . . . . . . . . . . 10
2.2.2 Extraction/propagation  . . . . . . . . 10
2.2.3 Transformation/cleansing  . . . . . . 10
2.2.4 Data refining  . . . . . . . . . . . . . . . . 11
2.2.5 Datamarts  . . . . . . . . . . . . . . . . . . 12
2.2.6 Metadata  . . . . . . . . . . . . . . . . . . . 12
2.2.7 Operational Data Store (ODS)  . . 15
2.3 Analytical users rEQuirements  . . . 16
2.3.1 Reporting and query  . . . . . . . . . . 17
2.3.2 On-Line Analytical Processing (OLAP) . . . . . . . . . . . . . . . . . . . . . . . 17
2.3.4 Statistics  . . . . . . . . . . . . . . . . . . . 21
2.3.5 Data mining  . . . . . . . . . . . . . . . . . 21
2.4 Data warehouse, OLAP and data mining summary. . . . . . . . . . . . . . . 21
Chapter 3. A generic data mining method . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1 What is data mining?  . . . . . . . . . . 24
3.2 What is new wITh data mining?  . . . 25
3.3 Data mining techniques  . . . . . . . . 27
3.3.1 Types of techniques  . . . . . . . . . . 27
3.3.2 Different applications that data mining can be used for . . . . . . . . . . 28
3.4 The generic data mining method  . 29
3.4.1 Step 1 - Defining the business issue. . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4.2 Step 2 - Defining a data model to use. . . . . . . . . . . . . . . . . . . . . . . . 35
3.4.3 Step 3 - Sourcing and preprocessing the data . . . . . . . . . . . . . . . . . 37
3.4.4 Step 4 - Evaluating the data model . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4.5 Step 5 -Choosing the data mining technique . . . . . . . . . . . . . . . . . . 41
3.4.6 Step 6 - IntERPreting the results  . . 42
3.4.7 Step 7 - Deploying the results  . . . 42
3.4.8 Skills rEQuired  . . . . . . . . . . . . . . . 43
3.4.9 Effort rEQuired  . . . . . . . . . . . . . . . 45
Chapter 4. How to discover the characteristics of your customers? . . . . 47
4.1 The business issue . . . . . . . . . . . . 48
4.1.1 How can data mining help? . . . . . 48
4.1.2 Where should we start?  . . . . . . . 49
4.2 The data to be used  . . . . . . . . . . . 51
4.2.1 Behavior data  . . . . . . . . . . . . . . . 51
4.2.2 Demographic data . . . . . . . . . . . . 52
4.2.3 AddITional data  . . . . . . . . . . . . . . 52
4.2.4 Data model for segmentation  . . . 53
4.3 Sourcing and preprocessing the data . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.1 Behavior data  . . . . . . . . . . . . . . . 55
4.3.2 Demographic data . . . . . . . . . . . . 55
4.4 Evaluating the data  . . . . . . . . . . . 56
4.5 The mining techniques  . . . . . . . . . 59
4.5.1 Choosing the mining techniques  . 59
4.5.2 Applying the mining techniques  . 60
4.6 IntERPreting the results  . . . . . . . . . 64
4.6.1 How to read and evaluate the results . . . . . . . . . . . . . . . . . . . . . . . . 64
4.6.2 IntERPreting the results from business perspective . . . . . . . . . . . . . . 67
4.7 Deploying the mining results  . . . . 73
4.7.1 Model deployment in various applications . . . . . . . . . . . . . . . . . . . . 73
4.7.2 Model deployment in marketing campaign process . . . . . . . . . . . . . 73
4.7.3 Model maintenance . . . . . . . . . . . 74
Chapter 5. Can you predict the customers who are likely to leave? . . . . 75
5.1 The business issue . . . . . . . . . . . . 76
5.1.1 How can data mining help? . . . . . 76
5.1.2 Where should we start?  . . . . . . . 76
5.1.3 Churn definITion  . . . . . . . . . . . . . . 77
5.1.4 Churn filtering  . . . . . . . . . . . . . . . 78
5.2 The data to be used  . . . . . . . . . . . 79
5.2.1 Churn indicator  . . . . . . . . . . . . . . 79
5.2.2 Customer information data  . . . . . 79
5.2.3 Call data  . . . . . . . . . . . . . . . . . . . 80
5.2.4 billing and payment data . . . . . . . 81
5.2.5 Key indices derived from call transaction data . . . . . . . . . . . . . . . . . 81
5.2.6 AddITional data  . . . . . . . . . . . . . . 81
5.2.7 Data model for churn prediction  . 81
5.3 Sourcing and preprocessing the data . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.3.1 Churn indicator  . . . . . . . . . . . . . . 87
5.3.2 Customer information data  . . . . . 87
5.3.3 Call data  . . . . . . . . . . . . . . . . . . . 87
5.3.4 billing and payment data . . . . . . . 88
5.3.5 Key indices derived from transaction data . . . . . . . . . . . . . . . . . . . . 88
5.4 Evaluating the data  . . . . . . . . . . . 89
5.5 The mining technique  . . . . . . . . . . 92
5.5.1 Choosing the mining technique . . 92
5.5.2 Applying the mining technique  . . 93
5.6 IntERPreting the results  . . . . . . . . . 97
5.6.1 IntERPreting results from business perspective . . . . . . . . . . . . . . . . . 97
5.6.2 Performance comparison  . . . . . 105
5.7 Deploying the model . . . . . . . . . . 108
5.7.1 Model deployment in various applications . . . . . . . . . . . . . . . . . . . 108
5.7.2 Model deployment in retention campaign process . . . . . . . . . . . . . 109
5.7.3 Model maintenance . . . . . . . . . . 109
Chapter 6. How to discover true value of your customers? . . . . . . . . . . 111
6.1 The business issue . . . . . . . . . . . 112
6.1.1 How can data mining help? . . . . 112
6.1.2 Where should we start?  . . . . . . 112
6.2 The data to be used  . . . . . . . . . . 114
6.2.1 Financial data  . . . . . . . . . . . . . . 114
6.2.2 CredIT risk  . . . . . . . . . . . . . . . . . 114
6.2.3 Behavior measure  . . . . . . . . . . . 114
6.2.4 Data used  . . . . . . . . . . . . . . . . . 115
6.3 Sourcing and preprocessing the data . . . . . . . . . . . . . . . . . . . . . . . . 116
6.3.1 Financial data  . . . . . . . . . . . . . . 116
6.3.2 CredIT risk  . . . . . . . . . . . . . . . . . 116
6.3.3 Behavior measure  . . . . . . . . . . . 116
6.4 Evaluating the data  . . . . . . . . . . 118
6.4.1 CredIT risk prediction  . . . . . . . . . 118
6.4.2 Customer value function (CVF)  . 121
6.5 The mining technique(s)  . . . . . . . 122
6.5.1 Choosing mining technique for credIT risk prediction. . . . . . . . . . . . 122
6.5.2 Choosing mining technique for CVF . . . . . . . . . . . . . . . . . . . . . . . . 122
6.6 IntERPreting the results  . . . . . . . . 124
6.6.1 CredIT risk prediction  . . . . . . . . . 124
6.6.2 CVF  . . . . . . . . . . . . . . . . . . . . . . 129
6.7 Deploying the mining results  . . . 133
..............................