Applied Predictive Analytics

Learn the art and science of predictive analytics -- techniques that get results
 
Predictive analytics is what translates big data into meaningful, usable business information. Written by a leading expert in the field, this guide examines the science of the underlying algorithms as well as the principles and best practices that govern the art of predictive analytics. It clearly explains the theory behind predictive analytics, teaches the methods, principles, and techniques for conducting predictive analytics projects, and offers tips and tricks that are essential for successful predictive modeling. Hands-on examples and case studies are included.
* The ability to successfully apply predictive analytics enables businesses to effectively interpret big data; essential for competition today
* This guide teaches not only the principles of predictive analytics, but also how to apply them to achieve real, pragmatic solutions
* Explains methods, principles, and techniques for conducting predictive analytics projects from start to finish
* Illustrates each technique with hands-on examples and includes as series of in-depth case studies that apply predictive analytics to common business scenarios
* A companion website provides all the data sets used to generate the examples as well as a free trial version of software
 
Applied Predictive Analytics arms data and business analysts and business managers with the tools they need to interpret and capitalize on big data.

List of contents

Introduction xxi
 
Chapter 1 Overview of Predictive Analytics 1
 
What Is Analytics? 3
 
What Is Predictive Analytics? 3
 
Supervised vs. Unsupervised Learning 5
 
Parametric vs. Non-Parametric Models 6
 
Business Intelligence 6
 
Predictive Analytics vs. Business Intelligence 8
 
Do Predictive Models Just State the Obvious? 9
 
Similarities between Business Intelligence and Predictive Analytics 9
 
Predictive Analytics vs. Statistics 10
 
Statistics and Analytics 11
 
Predictive Analytics and Statistics Contrasted 12
 

Predictive Analytics vs. Data Mining 13
 
Who Uses Predictive Analytics? 13
 
Challenges in Using Predictive Analytics 14
 
Obstacles in Management 14
 
Obstacles with Data 14
 
Obstacles with Modeling 15
 
Obstacles in Deployment 16
 
What Educational Background Is Needed to Become a Predictive Modeler? 16
 
Chapter 2 Setting Up the Problem 19
 
Predictive Analytics Processing Steps: CRISP-DM 19
 
Business Understanding 21
 
The Three-Legged Stool 22
 
Business Objectives 23
 
Defining Data for Predictive Modeling 25
 
Defining the Columns as Measures 26
 
Defining the Unit of Analysis 27
 
Which Unit of Analysis? 28
 
Defining the Target Variable 29
 
Temporal Considerations for Target Variable 31
 
Defining Measures of Success for Predictive Models 32
 
Success Criteria for Classifi cation 32
 
Success Criteria for Estimation 33
 
Other Customized Success Criteria 33
 
Doing Predictive Modeling Out of Order 34
 
Building Models First 34
 
Early Model Deployment 35
 
Case Study: Recovering Lapsed Donors 35
 
Overview 36
 
Business Objectives 36
 
Data for the Competition 36
 
The Target Variables 36
 
Modeling Objectives 37
 
Model Selection and Evaluation Criteria 38
 
Model Deployment 39
 
Case Study: Fraud Detection 39
 
Overview 39
 
Business Objectives 39
 
Data for the Project 40
 
The Target Variables 40
 
Modeling Objectives 41
 
Model Selection and Evaluation Criteria 41
 
Model Deployment 41
 
Summary 42
 
Chapter 3 Data Understanding 43
 
What the Data Looks Like 44
 
Single Variable Summaries 44
 
Mean 45
 
Standard Deviation 45
 
The Normal Distribution 45
 
Uniform Distribution 46
 
Applying Simple Statistics in Data Understanding 47
 
Skewness 49
 
Kurtosis 51
 
Rank-Ordered Statistics 52
 
Categorical Variable Assessment 55
 
Data Visualization in One Dimension 58
 
Histograms 59
 
Multiple Variable Summaries 64
 
Hidden Value in Variable Interactions: Simpson's Paradox 64
 
The Combinatorial Explosion of Interactions 65
 
Correlations 66
 
Spurious Correlations 66
 
Back to Correlations 67
 
Crosstabs 68
 
Data Visualization, Two or Higher Dimensions 69
 
Scatterplots 69
 
Anscombe's Quartet 71
 
Scatterplot Matrices 75
 
Overlaying the Target Variable in Summary 76
 
Scatterplots in More Than Two Dimensions 78
 
The Value of Statistical Signifi cance 80
 
Pulling It All Together into a Data Audit 81
 
Summary 82
 
Chapter 4 Data Preparation 83
 
Variable Cleaning 84
 
Incorrect Values 84
 
Consistency in Data Formats 85
&nbs

About the author

DEAN ABBOTT is President of Abbott Analytics, Inc. (San Diego). He is an internationally recognized data mining and predictive analytics expert with over two decades experience in fraud detection, risk modeling, text mining, personality assessment, planned giving, toxicology, and other applications. He is also Chief Scientist of SmarterRemarketer, a company focusing on behaviorally- and data-driven marketing and web analytics.

Summary

Learn the art and science of predictive analytics -- techniques that get results

Predictive analytics is what translates big data into meaningful, usable business information. Written by a leading expert in the field, this guide examines the science of the underlying algorithms as well as the principles and best practices that govern the art of predictive analytics. It clearly explains the theory behind predictive analytics, teaches the methods, principles, and techniques for conducting predictive analytics projects, and offers tips and tricks that are essential for successful predictive modeling. Hands-on examples and case studies are included.
* The ability to successfully apply predictive analytics enables businesses to effectively interpret big data; essential for competition today
* This guide teaches not only the principles of predictive analytics, but also how to apply them to achieve real, pragmatic solutions
* Explains methods, principles, and techniques for conducting predictive analytics projects from start to finish
* Illustrates each technique with hands-on examples and includes as series of in-depth case studies that apply predictive analytics to common business scenarios
* A companion website provides all the data sets used to generate the examples as well as a free trial version of software

Applied Predictive Analytics arms data and business analysts and business managers with the tools they need to interpret and capitalize on big data.

Additional text

"This book provides an excellent background to predictive analytics" (BCS, December 2014)

Report

"This book provides an excellent background to predictive analytics" (BCS, December 2014)

Product details

• Authors: D Abbott, Dean Abbott, Abbott Dean
• Publisher: Wiley, John and Sons Ltd
• Languages: English
• Product format: Paperback / Softback
• Released: 23.05.2014
• EAN: 9781118727966
• ISBN: 978-1-118-72796-6
• No. of pages: 456
• Dimensions: 190 mm x 237 mm x 25 mm
• Subjects: Natural sciences, medicine, IT, technology > IT, data processing > IT; Informatik, Datenanalyse, computer science, Database & Data Warehousing Technologies, Datenbanken u. Data Warehousing