상세정보

The potential business advantages of data mining are well documented in publications for executives and managers. However, developers implementing major data-mining systems need concrete information about the underlying technical principles?and their practical manifestations?in order to either integrate commercially available tools or write data-mining programs from scratch. This book is the first technical guide to provide a complete, generalized roadmap for developing data-mining applications, together with advice on performing these large-scale, open-ended analyses for real-world data warehouses.

Note: If you already own Predictive Data Mining: A Practical Guide, please see ISBN 1-55860-477-4 to order the accompanying software. To order the book/software package, please see ISBN 1-55860-478-2.

Reviews

Feature

정보제공 :

CONTENTS

Preface = xi

1 What Is Data Mining? = 1

 1.1 Big Data = 2

  1.1.1 The Data Warehouse = 3

  1.1.2 Timelines = 6

 1.2 Types of Data-Mining Problems = 7

 1.3 The Pedigree of Data Mining = 11

  1.3.1 Databases = 11

  1.3.2 Statistics = 12

  1.3.3 Machine Learnig = 13

 1.4 Is Big Better? = 14

  1.4.1 Strong Statistical Evaluation = 14

  1.4.2 More Intensive Search = 14

  1.4.3 More Controlled Experiments = 15

  1.4.4 Is Big Necessary? = 15

 1.5 The Tasks of Predicitive Data Mining = 16

  1.5.1 Data Preparation = 16

  1.5.2 Data Reduction = 18

  1.5.3 Data Modeling and Prediction = 19

  1.5.4 Case and Solution Analyses = 19

 1.6 Data Mining : Art or Science? = 21

 1.7 An Overview of the Bok = 21

 1.8 Bibliographic and Historical Remarks = 22

2 Statistical Evalustion for Big Data = 25

 2.1 The Idealized Model = 26

  2.1.1 Classical Statistical Comparison and Evaluation = 27

 2.2 It's Big but Is It Biased? = 30

  2.2.1 Objective Versus Survey Data = 30

  2.2.2 Significance and Purvey Data = 30

   2.2.2.1 Too Many Comparisons? = 32

 2.3 Classical Types of Statistical Prediction = 33

  2.3.1 Predicting True-or-False : Classification = 34

   2.3.1.1 Error Rates = 34

  2.3.2 Forecasting Numbers : Regression = 34

   2.3.2.1 Distance Measures = 35

 2.4 Measuring Predictive Performance = 36

  2.4.1 Independent Testing = 36

   2.4.1.1 Random Training and Testing = 36

   2.4.1.2 How Accurate Is the Error Estimate? = 38

   2.4.1.3 Comparing Results for Error Measures = 39

   2.4.1.4 Ideal or Real-World Sampling? = 41

   2.4.1.5 Training and Testing from Different Time Periods = 43

 2.5 Too Much Searching and Testing? = 45

 2.6 Why Are Errors Made? = 47

 2.7 Bibliographic and Historical Remarks = 49

3. Prearing the Data = 51

 3.1 A Standard From = 52

  3.1.1 Standard Measurements = 53

  3.1.2 Goals = 55

 3.2 Data Transformations = 55

  3.2.1 Normalizations = 57

  3.2.2 Data Smoothing = 58

  3.2.3 Differences and Ratios = 60

 3.3 Missing Data = 61

 3.4 Time-Dependent Data = 62

  3.4.1 Composing Features from Time Series = 67

   3.4.2.1 Current Values = 68

   3.4.2.2 Moving Averages = 68

   3.4.2.3 Trends = 69

   3.4.2.4 Seasonal Adjustments = 70

 3.5 Hybrid Time-Dependent Applications = 71

  3.5.1 Multivariate Time Series = 72

  3.5.2 Classification and Time Series = 73

  3.5.3 Standard Cases with Time-Series Attributes = 73

 3.6 Text Mining = 74

 3.7 Bibliogrphic and Historical Remarks = 78

4 Data Reduction = 81

 4.1 Selecting the Best Features = 84

 4.2 Feature Selection From Means and Variances = 86

  4.2.1 Independent Features = 87

  4.2.2 Distance-Based Optimal Feature Selection = 88

  4.2.3 Heuristic Feature Selection = 90

 4.3 Principal Components = 92

 4.4 Feature Selection by Decision Trees = 95

 4.5 How Many Measured Values? = 96

  4.5.1 Reducing and Smothing Values = 98

   4.5.1.1 Rounding = 101

   4.5.1.2 K-Means Clustering = 102

   4.5.1.3 Class Entropy = 104

 4.6 How Many Cases? = 106

  4.6.1 A Single Sample = 109

  4.6.2 Incremental Samples = 111

  4.6.3 Average Samples = 113

  4.6.4 Specialized Case-Reduction Techniques = 115

   4.6.4.1 Sequential Sampling over Time = 115

   4.6.4.2 Strategic Sampling of Key Events = 116

   4.6.4.3 Adjusting Prevalence = 116

 4.7 Bibliographic and Historical Remarks = 117

5 Looking for Solutions = 119

 5.1 Overview = 119

 5.2 Math Solutions = 120

  5.2.1 Linear Scoring = 120

  5.2.2 Nonlinear Scoring : Neural Nets = 123

  5.2.3 Advanced Statistical Methods = 128

 5.3 Distance Solutions = 132

 5.4 Logic Solutions = 135

  5.4.1 Decision Trees = 136

  5.4.2 Decision Rules = 138

 5.5 What Do the Answers Mean? = 142

  5.5.1 Is It Safe to Edit Solutions? = 144

 5.6 Which Solution Is Preferable? = 145

 5.7 Combining Differenct Answers = 146

  5.7.1 Multiple Prediction Methods = 147

  5.7.2 Multiple Samples = 148

 5.8 Bibliographic and Historical Reamrks = 150

6 What's Best for Data Reduction and Mining? = 153

 6.1 Let's Analyze Some Real Data = 154

 6.2 The Experimental Methods = 158

 6.3 The Empirical Results = 161

  6.3.1 Significance Testing = 162

 6.4 So What Did We Learn? = 162

  6.4.1 Feature Selection = 163

  6.4.2 Value Reduction = 167

  6.4.3 Subsampling or All Cases? = 170

 6.5 Graphical Trend Analysis = 172

  6.5.1 Incremental Case Analysis = 173

  6.5.2 Incremental Complexity Analysis = 176

 6.6 Maximum Data Reduction = 181

 6.7 Are There Winners and Losers in Performance? = 182

 6.8 Getting the Best Results = 184

 6.9 Bibliographic and Historical Remarks = 187

7 Art ro Science? Case Studies in Data Mining = 189

 7.1 Why These Case Studies? = 190

 7.2 A Summary of Tasks for Predictive Data Mining = 191

  7.2.1 A Checklist Data Preparation = 192

  7.2.2 A Checklist Data Reduction = 192

  7.2.3 A Checklist Data Modeling and Prediction = 192

  7.2.4 A Checklist Case and Solution Analyses = 193

 7.3 The Case Studies = 193

  7.3.1 Transaction Processing = 193

  7.3.2 Text Mining = 197

  7.3.3 Outcomes Analysis = 199

  7.3.4 Process Control = 202

  7.3.5 Marketing and User Profiling = 205

  7.3.6 Exploratory Analysis = 207

 7.4 Looking Ahead = 210

 7.5 Bibliographic and Historical Remarks = 211

Appendix : Data-Miner Software Kit = 213

References = 215

Author Index = 223

Subject Index = 225