Handbook of Statistical Analysis

AI and ML Applications

3rd Edition - September 16, 2024
Authors: Robert Nisbet, Gary D. Miner, Keith McCormick
Language: English
Hardback ISBN:
9 7 8 - 0 - 4 4 3 - 1 3 2 7 3 - 5
Paperback ISBN:
9 7 8 - 0 - 4 4 3 - 1 5 8 4 5 - 2
eBook ISBN:
9 7 8 - 0 - 4 4 3 - 1 5 8 4 6 - 9

Handbook of Statistical Analysis: AI and ML Applications, third edition, is a comprehensive introduction to all stages of data analysis, data preparation, model building, and model… Read more

Purchase options

Institutional subscription on ScienceDirect

Request a sales quote

Resources

Companion materials(opens in new tab/window)Textbook support for instructors(opens in new tab/window)

Handbook of Statistical Analysis: AI and ML Applications, third edition, is a comprehensive introduction to all stages of data analysis, data preparation, model building, and model evaluation. This valuable resource is useful to students and professionals across a variety of fields and settings: business analysts, scientists, engineers, and researchers in academia and industry. General descriptions of algorithms together with case studies help readers understand technical and business problems, weigh the strengths and weaknesses of modern data analysis algorithms, and employ the right analytical methods for practical application.

This resource is an ideal guide for users who want to address massive and complex datasets with many standard analytical approaches and be able to evaluate analyses and solutions objectively. It includes clear, intuitive explanations of the principles and tools for solving problems using modern analytic techniques; offers accessible tutorials; and discusses their application to real-world problems.

Cover image
Title page
Table of Contents
Copyright
Foreword 1 for first edition
Foreword 2 for first edition
Foreword—to this third edition
Preface
Overall organization of this book
Acknowledgments
Praise for the first edition of this book
Biographies of the primary authors of this book
Part I: Introduction
1. The background and history of predictive analytics
Abstract
Preamble
Phases in the history of statistics and predictive analytics
Current trends of development in predictive analytics
Postscript
References
2. Theoretical considerations for data analytics
Abstract
Preamble
The scientific method
What is the process of data analytics?
A theoretical framework for the data analytics process
Strengths of the data analytics process
Customer centric versus account centric: a new way to look at your data
The data paradigm shift
Creation of the CAR
Major activities of predictive analytics
Major challenges of data analytics and predictive analytics
General examples of predictive analytics applications
Major issues in data analytics
General requirements for success in a data analytics project
Example of a data analytics project: classify a bat’s species by its sound
The importance of domain knowledge
Postscript
References
3. The data analytics process
Abstract
Preamble
Two approaches to understanding and problem solving
The role of process in data analytics
Business understanding phase (mostly art)
Data understanding phase (mostly science)
Data preparation phase (a mixture of art and science)
Modeling (a mixture of art and science)
Deployment (mostly art)
Closing the information loop* (art)
The art of data analytics
Postscript
References
4. Tools
Abstract
Prelude
Accessory tools
Pure coding languages
Menu-based procedural platforms
Multipersona data science and machine learning (DSML) platforms with graphical programming canvasses and AutoML capabilities
Advice on tool selection and professional development
Postscript
Reference
Part II: Data preparation
5. Data access
Abstract
Preamble
Knowledge from data: predictive analytics
Overall structure of the database – data modeling
Account-centric format of data in early computer databases
The customer-centric analytical data mart
The ETL process
Issues in data mart design
Current trends in data access and data structure design
References
6. Data Understanding
Abstract
Prelude
Four Data Understanding tasks in CRISP-DM
KNIME examples in Data Understanding
IBM SPSS modeler examples
Reference
7. Visualizing data for understanding
Abstract
Foundations of data visualization: John Tukey
Importance of the audience of visualizations
Seeing is believing
Case Study: visualizations of the Ames Housing data set
Postscript
References
8. Data cleaning
Abstract
Preamble
Data cleaning operation classification
Types of data cleaning operations
Technical correction of data
String manipulation
Data filtering
Time-series filtering
Outlier handling
Postscript
References
9. Data conditioning
Abstract
Preamble
Data conditioning operations that generate consistent data sets
Three types of data conditioning operations are discussed in this chapter
Postscript
References
10. Data and feature engineering
Abstract
Prelude
Definitions
Topics of discussion in Chapter 10
References
11. Feature selection
Abstract
Preamble
Modeling and sculpting
Some introductory concepts
Feature selection versus dimensionality reduction
Classification of dimensionality reduction methods
Decision tree processing elements useful as feature reduction filters
KNIME examples of feature selection methods
Correlation filtering (a filter method)
References
Additional bibliography for feature selection
12. A data preparation cookbook
Abstract
Preamble
Introduction
1 CRISP-DM phase I: business understanding
2 CRISP-DM phase II: data understanding
3 CRISP-DM phase III: data preparation
4 CRISP-DM phase IV: modeling
18 common mistakes in data preparation in predictive analytics projects
References
Part III: Modeling
13. A taxonomy of algorithms
Abstract
Preamble
Introduction
The top five modeling algorithms
More advanced algorithms
A brief introduction to the top five algorithms
Keywords and phrases
References
14. Modeling methods
Abstract
Prelude
Scope of modeling operations
General modeling issues
Types of analytical models
Two approaches to building analytical models
Two basic types of predictive models
Keywords and phrases
References
15. Model evaluation and enhancement
Abstract
Preamble
Evaluation and enhancement: part of the modeling process
Types of errors in analytical models
Assessment of random error
Model enhancement techniques
Model enhancement checklist
References
16. Ensembles and complexity: a practical approach
Abstract
Preamble
Modeling: How to choose the best modeling option?
Do ensembles violate Occam’s razor?
The power of ensembles
Types of ensemble modeling techniques
General structure of a fused ensemble model
References
17. Deep learning versus traditional machine learning
Abstract
Preamble
Introduction
Differences between deep learning and traditional machine learning
The journey to deep learning
Why is feature engineering so important
Interactions
Interaction representation in artificial neural networks
Postscript
References
18. Interpretable machine learning and explainable AI
Abstract
Preamble
Introduction
Comparing interpretable machine learning and XAI
Comparing two types of XAI: global and local explanations
Techniques for global explanations
Techniques for local explanations
LIME
Regulatory considerations
The proposed AI Act in the EU
Postscript
References
19. Human in the loop machine learning and data annotation
Abstract
Preamble
Introduction
Defining human in the loop
Concurrent rise of human in the loop and deep learning
Common human in the loop use cases
The data annotation ecosystem
The ethics of data annotation
Issues associated with managing a data annotation project
Becoming “data centric”
Postscript
References
Part IV: Applications
20. Overview of healthcare delivery and medical informatics
Abstract
Prelude
Introduction
Chapter 1: The purpose of the book on medical data analytics
Chapter 2: History of predictive analytics in medicine and health care
Chapter 3: Bioinformatics
Chapter 4: Data and process models in medical informatics
Chapter 5: Access to data for analytics – the initial issue in medical and healthcare predictive analytics
Chapter 6: Precision (personalized) medicine
Chapter 7: Patient-directed health care
Chapter 8: Regulatory measures – agencies and data issues in medicine and health care
Chapter 9: Predictive analytics with multiomics data
Chapter 10: Artificial intelligence and genomics
Chapter 11: Glaucoma (eye disease): a case study with suggested predictive analytics modeling for identifying an individual patient’s best diagnosis and best treatment
Chapter 12: Practical application example: using data science algorithms in predicting ICU patient urine output in response to diuretics to aid clinicians and healthcare providers in clinical decision making
Chapter 13: Practical application example: prediction tool development
Chapter 14: Modeling precancerous colon polyps with OMOP data
Chapter 15: Prediction of pancreatic and lung cancer from metabolomics data
Chapter 16: COVID-19 descriptive analytics visualization of pandemic and hospitalization data
Chapter 17: Disseminated intravascular coagulation predictive analytics with pediatric ICU admissions
Chapter 18: Challenges for healthcare administration and delivery: integrating predictive and prescriptive modeling into personalized precision health care
Chapter 19: Challenges of medical research in incorporating modern data analytics in studies
Chapter 20: The nature of insight from data and implications for automated decisioning
Chapter 21: Model management and ModelOps: managing an artificial intelligence-driven enterprise
Chapter 22: Forecasts for advances in predictive and prescriptive analytics
Chapter 23: Sampling and data analysis
Chapter 24: Analytics architectures for the 21st century
Chapter: 25 Predictive models versus prescriptive models; causal inference and Bayesian networks
Chapter 26: The future: 21st-century health care and wellness in the digital age
Postscript
References
21. Customer response modeling with temporal abstractions
Abstract
Preamble
Early CRM issues in business
Knowing how customers behaved before they acted
Transforming corporations into business ecosystems: The path to customer fulfillment
CRM modeling in business ecosystems
Modeling with different sets of predictor variables
Discussion
Conclusions
Postscript
References
22. Academic Analytics: a case study in student retention
Abstract
Preamble
Introduction
Academic Analytics
Student retention: a case study
Methodology
Results
Key word list
Discussion
Postscript
References
Key word list
23. Medical informatics case study: predicting type 1 precancerous colon polyps
Abstract
Prelude
Introduction
The University of California, Irvine Colon Polyp Risk Modeling Project
Modeling
Results
Discussion
Postscript
References
24. Credit risk case study
Abstract
Prelude
Introduction
Credit scoring
Credit risk models
Case study: credit risk score for a loan application
Postscript
Reference
25. The intelligence revolution
Abstract
Preamble
Introduction
Large language models (ChatGPT, BARD, etc.)
TinyML
HUMANE
Tiny to Large AI-FAB LABS
Data and Predictive Analytics as a “Personal AI” or “Company AI” that functions automatically with AI and AGI – the future is only 3–7 years away!!!
Postscript
References
Part V: Right models – luck & ethics of analytics
26. The “Right Model” for the “Right Purpose” – when less is good enough
ABSTRACT
Preamble
More is not necessarily better: Lessons from nature and engineering
References
27. Ethics and data analytics
ABSTRACT
Preamble
The birthday party – a practical example for ethical action
Secular ethics
Michael Sandel on “Doing the Right Thing” with data analytics
Alignment of ethical perspectives: The ethical road map
What is generative artificial intelligence?
Postscript
References
28. The significance of luck
ABSTRACT
Preamble
Introduction
The problem of significance in traditional p-value statistical analysis
The response of the American Statistical Association
Postscript
References
Part VI: Tutorials and case studies
Tutorial A. Example of data mining recipes using Windows 10 and STATISTICA 13
Open the data set
Open the Data Miner Recipes interface
Opening the data set
Fixing the text label problem
Data mining recipes (again)
Tutorial B. Analysis of Hurricane data (Hurrdata.sta) using the STATISTICA Data Miner 13
Discussion
Tutorial C. Case study – Predicting student success at high-stakes nursing examinations (NCLEX) using IBM SPSS Modeler and STATISTICA Data Miner 13
Introduction
Decision management in nursing education
Case study
Research question
Literature review
Dataset and expected strength of predictors
Data mining with IBM SPSS Modeler
Improving model accuracy and stability: boosting and bagging
Data mining with STATISTICA
Conclusion
References
Uncited references for Additional Information
Tutorial D. Constructing a Histogram using MidWest Company Personality Data with KNIME
Changing to the KNIME classical interface
Opening KNIME
Tutorial E. Feature selection using KNIME
Why select features?
Occam’s razor – simple, but not simplistic
Local minimum error
Moving out of the local minimum
Strategies for reduction of dimensionality in predictive analytics available in KNIME
Tutorial F. Medical/business tutorial using STATISTICA Data Miner 13 determining possible predictors for days with hospice service for patients with dementia
Tutorial G. Modeling hospice residence stay with KNIME
Introduction
Modeling objective
Data
Adding an extension to KNIME
Creating the Tutorial G workflow
The XGBoost Tree Ensemble modeling algorithm
Model evaluation
Discussion of model accuracy results
Feature importance
Tutorial H. Data preparation: Merging data sources
Tutorial I. Data description
Tutorial J. Data cleaning and recoding
Tutorial K. Dummy coding category variables
Tutorial L. Outlier handling
Tutorial M. Filling missing values with constants
Tutorial N. Filling missing values with formulas
Tutorial O. Filling missing values with a model
APPENDIX A. Listing of tutorials and other resources on this book’s companion web page
Appendix B
Index

Robert Nisbet

Bob Nisbet, PhD, is a Data Scientist, currently modeling precancerous colon polyp presence with clinical data at the UC-Irvine Medical Center. He has experience in predictive modeling in Telecommunications, Insurance, Credit, Banking. His academic experience includes teaching in Ecology and in Data Science. His industrial experience includes predictive modeling at AT&T, NCR, and FICO. He has worked also in Insurance, Credit, membership organizations (e.g. AAA), Education, and Health Care industries. He retired as an Assistant Vice President of Santa Barbara Bank & Trust in charge of business intelligence reporting and customer relationship management (CRM) modeling.

Affiliations and expertise

Researcher-Medical Informatics, H.H. Chao Comprehensive Digestive Disease Center, University of California Irvine Medical Center, Private Consulting, Santa Barbara, CA, USA

Gary D. Miner

Dr. Gary Miner PhD received a B.S. from Hamline University, St. Paul, MN, with biology, chemistry, and education majors; an M.S. in zoology and population genetics from the University of Wyoming; and a Ph.D. in biochemical genetics from the University of Kansas as the recipient of a NASA pre-doctoral fellowship. He pursued additional National Institutes of Health postdoctoral studies at the U of Minnesota and U of Iowa eventually becoming immersed in the study of affective disorders and Alzheimer's disease. In 1985, he and his wife, Dr. Linda Winters-Miner, founded the Familial Alzheimer's Disease Research Foundation, which became a leading force in organizing both local and international scientific meetings, bringing together all the leaders in the field of genetics of Alzheimer's from several countries, resulting in the first major book on the genetics of Alzheimer’s disease. In the mid-1990s, Dr. Miner turned his data analysis interests to the business world, joining the team at StatSoft and deciding to specialize in data mining. He started developing what eventually became the Handbook of Statistical Analysis and Data Mining Applications (co-authored with Drs. Robert A. Nisbet and John Elder), which received the 2009 American Publishers Award for Professional and Scholarly Excellence (PROSE). Their follow-up collaboration, Practical Text Mining and Statistical Analysis for Non-structured Text Data Applications, also received a PROSE award in February of 2013. Gary was also co-author of “Practical Predictive Analytics and Decisioning Systems for Medicine (Academic Press, 2015). Overall, Dr. Miner’s career has focused on medicine and health issues, and the use of data analytics (statistics and predictive analytics) in analyzing medical data to decipher fact from fiction. Gary has also served as Merit Reviewer for PCORI (Patient Centered Outcomes Research Institute) that awards grants for predictive analytics research into the comparative effectiveness and heterogeneous treatment effects of medical interventions including drugs among different genetic groups of patients; additionally he teaches on-line classes in ‘Introduction to Predictive Analytics’, ‘Text Analytics’, ‘Risk Analytics’, and ‘Healthcare Predictive Analytics’ for the University of California-Irvine. Recently, until ‘official retirement’ 18 months ago, he spent most of his time in his primary role as Senior Analyst-Healthcare Applications Specialist for Dell | Information Management Group, Dell Software (through Dell’s acquisition of StatSoft (www.StatSoft.com) in April 2014). Currently Gary is working on two new short popular books on ‘Healthcare Solutions for the USA’ and ‘Patient-Doctor Genomics Stories’.

Affiliations and expertise

CEO, M&M Predictive Analytics LLC; UCI Adjunct Professor for Continuing Education, Predictive Analytics Program; Associate Editor, The Journal of Geriatric Psychiatry and Neurology; Private Consulting, Tulsa, OK, USA

Keith McCormick

Keith McCormick is a highly accomplished professional consultant, mentor, and trainer, having served as keynote and moderator at international conferences focused on analytic practitioners and leadership alike. Keith has leveraged statistical software since 1990 along with deep expertise utilizing popular industry advanced analytics solutions such as IBM SPSS Statistics, IBM SPSS Modeler, KNIME, popular open-source and other tools involving text and big data analytics. Keith is currently Data Science Principal with Further.

Affiliations and expertise

Data Science Principal, Further

Life Sciences

Physical Sciences & Engineering

Social Sciences & Humanities

Health

Handbook of Statistical Analysis

AI and ML Applications

Purchase options

Institutional subscription on ScienceDirect

Resources

Robert Nisbet

Gary D. Miner

Keith McCormick