"A good stack of examples, as large as possible, is indispensable for a thorough understanding of any concept, and when I want to learn something new, I make it my first job to build one." – Paul Halmos, mathematician (by coincidence, saw this quotation twice today, but it is a big reason to read this book, as you will see).
This is my new favourite applied mathematics book, the equal of Misner, Thorne and Wheeler's Gravitation (1971).
If you want to write technical books, it is worth getting the free sample chapters of this book, even if you’re not interested in the topics, just to see how to write books on difficult technical topics, and really worth getting if you are interested in statistics and probability. Even if the subject matter is familiar, there are so many good explanations and so many examples for each topic, drawn from different disciplines, that it is worth reading just you can use with your less informed friends.
I wanted to learn Bayesian Networks. I bought 3 books. This one, Causality, by Judea Pearl and Bayesian Networks in R with Applications in Systems Biology by Rahakrishnan Nagarajan et. al. The first is also very good; review later as I'm not so far into it. The other is awful, I'll review that too. “Causality” revealed some holes in my memory of basic probability and statistics, which is why I bought “Risk Assessment...” (RA). I also got a bunch of Internet resources because when learning without an advisor, I need help in case I don’t understand something important. However, so far (half-way through), I’ve not needed anything else at all. Judea Pearl, who is a leading authority on these topics, wrote the very approving Foreword to this book and I’m now finding the more advance Pearl book much easier now I’ve read this.
RA is written in Tufte style, which I've adopted myself for my own writing. If you don’t know Edward Tufte’s writing on how to present numbers and facts, that's worth looking into, but you’ll get a really good demonstration from reading this book. The advantage is that the main text is supplemented by sidebars of illustrations, examples and further explanations so that you can either go fast where you follow the concepts or look at the sidebars for more explanation if you get bogged down. And even though the book eases you into each topic, you also get the formal mathematical notations and proofs so you can understand related papers and books. And you can use the many examples as exercises. Stop and work them yourself before reading on.
All this, plus access to free software and data to work the examples that need computer assistance. Although I’m doing working in Python and instead, the AgenaRisk software is much more user friendly. If you are not a programmer and don’t want to be, you can still use a computer to to see what is going on. If you know R or Python, you can easily The book web site (http://www.bayesianrisk.com/) links to the free chapters and to a blog and other supporting material.
I wish I’d had this 10 years ago when risk assessment was a big part of my job; I just got this because 90% of the book is setting up the probability and statistical machinery so even if not interested in risk management, but it I have read the risk parts and they are really good for anyone running a more-than-tiny project where things can go wrong. Risk management is a huge part of any project management. If you’re still at the “risk=probability×impact” stage, this book will up your game a lot.
Wednesday 28 March 2018
We found this review on goodreads.com
There is a review of our book in the Journal of the ICACA (full pdf of the review is here)
Start at the beginning, and you will find it is clear that this book is written to be understandable by professional people who are interested in risk assessment and decision making; readers do not need an in-depth knowledge of statistics for the book to be enjoyable and useful.We can even forgive the reference to "Agenda Risk" rather than "AgenaRisk"
For those looking to begin working with Bayesian networks, this book serves as an excellent starting point and provides guidance for readers on how to develop and run a Bayesian network model for risk assessment and decision making.