Dahan / Cohen / Rokach Proactive Data Mining with Decision Trees

1. Auflage 2014
ISBN: 978-1-4939-0539-3
Verlag: Springer
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 94 Seiten

Reihe: SpringerBriefs in Electrical and Computer Engineering

This book explores a proactive and domain-driven method to classification tasks. This novel proactive approach to data mining not only induces a model for predicting or explaining a phenomenon, but also utilizes specific problem/domain knowledge to suggest specific actions to achieve optimal changes in the value of the target attribute. In particular, the authors suggest a specific implementation of the domain-driven proactive approach for classification trees. The book centers on the core idea of moving observations from one branch of the tree to another. It introduces a novel splitting criterion for decision trees, termed maximal-utility, which maximizes the potential for enhancing profitability in the output tree. Two real-world case studies, one of a leading wireless operator and the other of a major security company, are also included and demonstrate how applying the proactive approach to classification tasks can solve business problems. Proactive Data Mining with Decision Trees is intended for researchers, practitioners and advanced-level students.

Autoren/Hrsg.

Dahan, Haim

Cohen, Shahar

Rokach, Lior

Maimon, Oded

Inhaltsverzeichnis

1;Preface;6
2;Contents;8
3;Chapter 1 Introduction to Proactive Data Mining;10
3.1;1.1 Data Mining;10
3.2;1.2 Classification Tasks;10
3.3;1.3 Basic Terms;11
3.3.1;1.3.1 Training Set;11
3.3.2;1.3.2 Classification Task;11
3.3.3;1.3.3 Induction Algorithm;12
3.4;1.4 Decision Trees (Classification Trees);12
3.5;1.5 Cost Sensitive Classification Trees;15
3.6;1.6 Classification Trees Limitations;17
3.7;1.7 Active Learning;17
3.8;1.8 Actionable Data Mining;19
3.9;1.9 Human Cooperated Mining;20
3.10;References;21
4;Chapter 2 Proactive Data Mining: A General Approach and Algorithmic Framework;24
4.1;2.1 Notations;24
4.2;2.2 From Passive to Proactive Data Mining;25
4.3;2.3 Changing the Input Data;26
4.4;2.4 The Need for Domain Knowledge: Attribute Changing Cost and Benefit Functions;27
4.5;2.5 Maximal Utility: The Objective of Proactive Data Mining Tasks;27
4.6;2.6 An Algorithmic Framework for Proactive Data Mining;28
4.7;2.7 Chapter Summary;29
4.8;References;29
5;Chapter 3 Proactive Data Mining Using Decision Trees;30
5.1;3.1 Why Decision Trees?;30
5.2;3.2 The Utility Measure of Proactive Decision Trees;31
5.3;3.3 An Optimization Algorithm for Proactive Decision Trees;35
5.4;3.4 The Maximal-Utility Splitting Criterion;36
5.5;3.5 Chapter Summary;40
5.6;References;42
6;Chapter 4 Proactive Data Mining in the Real World: Case Studies;43
6.1;4.1 Proactive Data Mining in a Cellular Service Provider;43
6.1.1;4.1.1 The Data Mining Problem for the Wireless Company;43
6.1.2;4.1.2 The Wireless Dataset;44
6.1.3;4.1.3 Attribute Discretization;45
6.1.4;4.1.4 Additional Environment and Problem Knowledge for the Wireless Company;46
6.1.4.1;4.1.4.1 Cost Matrices;47
6.1.4.2;4.1.4.2 Benefit Matrix;49
6.1.5;4.1.5 Passive Classification Model for the Wireless Company;50
6.1.6;4.1.6 Maximal Utility Generated Model for the Wireless Company;51
6.1.7;4.1.7 Optimization Algorithm over the J48 Generated Model for the Wireless Company;54
6.1.8;4.1.8 Optimization Algorithm over the Maximal Utility Generated Model for the Wireless Company;54
6.2;4.2 The Security Company Case;56
6.2.1;4.2.1 The Data Mining Problem for the Security Company;57
6.2.2;4.2.2 The Security Dataset;58
6.2.3;4.2.3 Attribute Discretization;59
6.2.4;4.2.4 Additional Environment and Problem Knowledge for the Security Company;60
6.2.4.1;4.2.4.1 Cost Matrices;60
6.2.4.2;4.2.4.2 Benefit Matrix;62
6.2.5;4.2.5 Passive Classification Model for the Security Company;63
6.2.6;4.2.6 Maximal Utility Generated Model for the Security Company;63
6.2.7;4.2.7 Optimization Algorithm over the J48 Generated Model for the Security Company;64
6.2.8;4.2.8 Optimization Algorithm over the Maximal Utility Generated Model for the Security Company;66
6.3;4.3 Case Studies Summary;68
6.4;References;69
7;Chapter 5 Sensitivity Analysis of Proactive Data Mining;70
7.1;5.1 Zero-one Benefit Function;70
7.2;5.2 Dynamic Benefit Function;76
7.3;5.3 Dynamic Benefits and Infinite Costs of the Unchangeable Attributes;78
7.4;5.4 Dynamic Benefit and Balanced Cost Functions;83
7.5;5.5 Chapter Summary;91
7.6;References;91
8;Chapter 6 Conclusions;93