A modern practical guide to building and using actuarial models. Loss From Data to Decisions is organized around the principle that actuaries build models in order to analyze risks and make decisions about managing the risks based on conclusions drawn from the analysis. In practice, one begins with data and ends with a business decision. The book flows logically from this principle. It begins with a framework for model building and a description of frequency and severity loss data typically available to actuaries. Parametric models are emphasized throughout. The frequency and severity models are used in building aggregate loss models, in credibility-based pricing models, and in loss analysis over multiple time periods. Designed as both an educational text as well as a professional reference, Loss
The book is definitely a very good and worthwhile read. It's clear the authors have a great understanding of how all the distributions work and how they're connected. However, this is definitely not the best choice for those who are unfamiliar with the subject. The book often explains one concept with the assumption that the reader is already quite familiar with several others. What's strange is that later, maybe 200 pages in, it will introduce those underlying concepts from the beginning as if the reader doesn't know them, which contradicts the earlier assumption of their understanding. So, while it's not an introductory-level book, it's still quite a good one for providing an overall understanding and showing the connections between various subjects. This is also one of the few books where the appendix is definitely worth paying attention to. What I didn't like was the generally intimidating feeling of the book. Even though the explanations are quite clear, maybe it's an editorial issue.
I read sections of this book more as an overview to loss models and credit risk analysis. It was informative for what I needed. Unfortunately most is for actuarial models and insurance. I would have liked to have seen more examples and some case studies and a bit less theory.