Kate Niswander recently talked with Glenn Meyers, co-editor of the newly released book, Predictive Modeling Applications in Actuarial Science, a two-part series published by Cambridge University Press and co-sponsored by the CAS. Volume One is now available for sale, with Volume Two to arrive in the fall of 2015.
How did this series come about?
Edward (Jed) Frees initiated the project and he invited Richard Derrig and me to join him as co-editors. Jed is the Hickman-Larson Professor of Actuarial Science at University of Wisconsin-Madison, and Richard Derrig is the president of Opal Consulting and visiting professor of risk, insurance and healthcare management at Temple University. I had been working with Jed on a predictive modeling project at ISO and we had written three joint papers (along with Dave Cummings, newly appointed CAS Vice President-Research and Development) on this project.
Who is the intended audience for Volume One?
This volume is for actuaries who wish to develop their expertise in statistics and to become familiar with concrete examples of predictive modeling. We hope the book will also address the needs of more seasoned practicing actuaries who would like an overview of advanced statistical topics that are relevant to actuarial practice.
How did you organize the book?
This series is unique because it covers a wide range of advanced statistical topics that are of interest to actuaries. With the distinctive chapter format, actuaries can use the book as a reference to zoom in on the topic of interest. Beginning with reviews of regression and time series methods, this book provides step-by-step introductions to advanced predictive modeling techniques that are particularly useful in actuarial practice. Readers will gain expertise in several statistical topics, including generalized linear modeling, and the analysis of longitudinal, two-part (frequency/severity) and fat-tailed data.
Which chapters do you find particularly relevant?
As readers of my Actuarial Review “Explorations” and “Brainstorms” columns will know, I have been using Markov chain Monte Carlo (MCMC) for a few years now to address a number of actuarial problems. The chapter “Bayesian Computational Methods,” by Brian Hartman, is a good introduction to the topic.
Was there anything about the book that surprised you?
In spite of my long career working on predictive modeling projects, this book covers a lot of predictive modeling topics I did not know—and some I wish I had known when working on these projects.
What tools and resources are available for those who purchase the book?
The book’s website is hosted by the University of Wisconsin and features many free resources for immediate use, including content preview, data (.txt or .csv format) and R code (R format) for each chapter. Members can access the website via the CAS press release for the book at www.casact.org.
Who worked with you on the book?
Several CAS members wrote chapters, including Gary Dean, Louise Francis and Jim Guszcza. Many CAS members also served as reviewers, including John Baldan, Lee Bowron, Dave Clark, Marc-André Desrosiers, Mario DiCaro, Rob Erhardt, Luyang Fu, Mark Goldburd, Anand Khare, Dan Tevet and Ben Walker.
What can CAS members take away from this book?
My advice to CAS members is to look closely at the table of contents and read through at least the introduction to each chapter to get a sense of the kind of problems that are being addressed. Then as you do your daily work in predictive modeling, dig deeper into the relevant chapters as needed.
What topics will we see in Volume Two?
In Volume One, my co-editors and I selected the chapter topics and recruited the authors with expertise in these topics. We also enlisted at least two experienced actuarial reviewers, including many CAS members, for each chapter. In Volume Two we approached various actuaries who are actively working in predictive modeling and asked them to write a chapter on one of their projects. Volume Two should be an interesting and expansive view of current practices in predictive modeling.
* * *
For more information and to order the book, please visit http://www.casact.org/PredictiveModelingBook. Proceeds from sales of this series will go to The Actuarial Foundation, a widely respected charitable organization that promotes actuarial scholarship, as well as its Canadian counterpart, the Actuarial Foundation of Canada.
Glenn Meyers, FCAS, MAAA, CERA, and Ph.D., retired from ISO after a 37 year career as an actuary.
Ms. Niswander joined the CAS office in 2014.