Practical Data Mining

294 Pages·2011·3.96 MB·English

by  Monte F. Hancock Jr

Information Technology/ Database H
a
Achieves a unique and delicate balance between depth, breadth, and clarity. n
c
—Stefan Joe-Yen, Cognitive Research Engineer, Northrop Grumman Corporation o
& Adjunct Professor, Department of Computer Science, Webster University c
k
Used as a primer for the recent graduate or as a refresher for the grizzled veteran,
Practical Data Mining is a must-have book for anyone in the field of data
mining and analytics.
P
—Chad Sessions, Program Manager, Advanced Analytics Group (AAG)
R
A
Used by corporations, industry, and government to inform and fuel everything from
C
focused advertising to homeland security, data mining can be a very useful tool
across a wide range of applications. Unfortunately, most books on the subject are
T
designed for the computer scientist and statistical illuminati and leave the reader
largely adrift in technical waters. I
C
Revealing the lessons known to the seasoned expert, yet rarely written down for
A
the uninitiated, Practical Data Mining explains the ins-and-outs of the detection,
characterization, and exploitation of actionable patterns in data. This working field L
manual outlines the what, when, why, and how of data mining and offers an easy-

to-follow, six-step spiral process.
D
Helping you avoid common mistakes, the book describes specific genres of data A
mining practice. Most chapters contain one or more case studies with detailed
T
project descriptions, methods used, challenges encountered, and results obtained.
The book includes working checklists for each phase of the data mining process. A
Your passport to successful technical and planning discussions with management,

senior scientists, and customers, these checklists lay out the right questions to ask
M
and the right points to make from an insider’s point of view.
I
Visit the book’s webpage for access to additional resources—including checklists,
N
figures, PowerPoint® slides, and a small set of simple prototype data mining tools.
I
http://www. celestech.com/PracticalDataMining N
G
K13109
ISBN: 978-1-4398-6836-2
90000
www.crcpress.com
9 781439 868362
www.auerbach-publications.com
K13109 cvr mech.indd 1 10/31/11 4:31 PM
Practical
Data
Mining
K13109_FM.indd 1 11/8/11 4:17 PM
TThhiiss ppaaggee iinntteennttiioonnaallllyy lleefftt bbllaannkk
Practical
Data
Mining
Monte F. Hancock, Jr.
Chief Scientist, Celestech, Inc.
K13109_FM.indd 3 11/8/11 4:17 PM
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2012 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
Version Date: 20111031
International Standard Book Number-13: 978-1-4398-6837-9 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts
have been made to publish reliable data and information, but the author and publisher cannot assume
responsibility for the validity of all materials or the consequences of their use. The authors and publishers
have attempted to trace the copyright holders of all material reproduced in this publication and apologize to
copyright holders if permission to publish in this form has not been obtained. If any copyright material has
not been acknowledged please write and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmit-
ted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying, microfilming, and recording, or in any information storage or retrieval system,
without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.
com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood
Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and
registration for a variety of users. For organizations that have been granted a photocopy license by the CCC,
a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used
only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Dedication
This book is dedicated to my beloved wife, Sandy, and to my dear little sister, Dr.
Angela Lobreto. You make life a joy.
Also, to my professional mentors George Milligan, Dr. Craig Price, and Tell Gates,
three of the finest men I have ever known, or ever hope to know: May God bless you
richly, gentlemen; He has blessed me richly through you.
v
TThhiiss ppaaggee iinntteennttiioonnaallllyy lleefftt bbllaannkk
Contents
Dedication v
Preface xv
About the Author xxi
Acknowledgments xxiii
Chapter 1 What Is Data Mining and What Can It Do? 1
Purpose 1
Goals 1
1.1 Introduction 1
1.2 A Brief Philosophical Discussion 2
1.3 The Most Important Attribute of the Successful
Data Miner: Integrity 3
1.4 What Does Data Mining Do? 4
1.5 What Do We Mean By Data? 6
1.5.1 Nominal Data vs. Numeric Data 7
1.5.2 Discrete Data vs. Continuous Data 7
1.5.3 Coding and Quantization as Inverse Processes 8
1.5.4 A Crucial Distinction: Data and Information Are
Not the Same Thing 9
1.5.5 The Parity Problem 11
1.5.6 Five Riddles about Information 11
1.5.7 Seven Riddles about Meaning 13
1.6 Data Complexity 14
1.7 Computational Complexity 15
vii
viii Practical Data Mining
1.7.1 Some NP-Hard Problems 17
1.7.2 Some Worst-Case Computational Complexities 17
1.8 Summary 17
Chapter 2 The Data Mining Process 19
Purpose 19
Goals 19
2.1 Introduction 19
2.2 Discovery and Exploitation 20
2.3 Eleven Key Principles of Information Driven Data Mining 23
2.4 Key Principles Expanded 24
2.5 Type of Models: Descriptive, Predictive, Forensic 30
2.5.1 Domain Ontologies as Models 30
2.5.2 Descriptive Models 32
2.5.3 Predictive Models 32
2.5.4 Forensic Models 32
2.6 Data Mining Methodologies 32
2.6.1 Conventional System Development:
Waterfall Process 33
2.6.2 Data Mining as Rapid Prototyping 34
2.7 A Generic Data Mining Process 34
2.8 RAD Skill Set Designators 35
2.9 Summary 36
Chapter 3 Problem Definition (Step 1) 37
Purpose 37
Goals 37
3.1 Introduction 37
3.2 Problem Definition Task 1: Characterize Your Problem 38
3.3 Problem Definition Checklist 38
3.3.1 Identify Previous Work 43
3.3.2 Data Demographics 45
3.3.3 User Interface 47
3.3.4 Covering Blind Spots 50
3.3.5 Evaluating Domain Expertise 51
3.3.6 Tools 53
3.3.7 Methodology 54
3.3.8 Needs 54
Contents ix
3.4 Candidate Solution Checklist 56
3.4.1 What Type of Data Mining Must the System
Perform? 56
3.4.2 Multifaceted Problems Demand Multifaceted
Solutions 57
3.4.3 The Nature of the Data 58
3.5 Problem Definition Task 2: Characterizing Your Solution 62
3.5.1 Candidate Solution Checklist 62
3.6 Problem Definition Case Study 64
3.6.1 Predictive Attrition Model: Summary Description 64
3.6.2 Glossary 64
3.6.3 The ATM Concept 65
3.6.4 Operational Functions 65
3.6.5 Predictive Modeling and ATM 67
3.6.6 Cognitive Systems and Predictive Modeling 68
3.6.7 The ATM Hybrid Cognitive Engine 68
3.6.8 Testing and Validation of Cognitive Systems 69
3.6.9 Spiral Development Methodology 69
3.7 Summary 70
Chapter 4 Data Evaluation (Step 2) 71
Purpose 71
Goals 71
4.1 Introduction 71
4.2 Data Accessibility Checklist 72
4.3 How Much Data Do You Need? 74
4.4 Data Staging 75
4.5 Methods Used for Data Evaluation 76
4.6 Data Evaluation Case Study: Estimating the
Information Content Features 77
4.7 Some Simple Data Evaluation Methods 81
4.8 Data Quality Checklist 85
4.9 Summary 87
Chapter 5 Feature Extraction and Enhancement (Step 3) 89
Purpose 89
Goals 89
5.1 Introduction: A Quick Tutorial on Feature Space 89
5.1.1 Data Preparation Guidelines 90