metode cluestring of data with a hierarchical cluestring

Although he was my great great uncle, I am not writing this review with bias. I found it to be a fascinating book, with a (relatively) new perspective on many matters

This book is written in a theorem-proof-corollary format and as such is not suitable to casual readers. However, to readers who want a deep understanding of the the theorems and proofs, it is a real gem. The material is very clear and simple (well at least simpler than other textbooks in this area). No prior knowledge is required! All the proofs are complete and succint.

Little did I expect of a book that has seemingly not garnered that much attention among professors and students (at least the lack of reviews in Amazon.com might serve as a confirming instance of that speculation), but come exam time, the book proved that such prejudice is outright foolish.

The authors must really love both the field and writing about it, for their overflowing exuberance readily transfers to the pages. Pictures and humor are never a scarcity here: the authors took no shame providing both; a curious but pleasurably fresh anomaly in the often dry and coldly serious world that is mathematics writing.

As a sample survey practitioner for over 25 years, I very highly recommend this book. It accomplishes what other books claim but fail, to be both for novices as well as the more experienced.

The author graciously understates his position as one of the luminaries of sampling, as evidenced by his professional publications from the 1960's, to his colleagues choosing him to write six entries (at last count) on various aspects of sample surveys in the Encyclopedia of Statistical Sciences. The author is a major contributor to the main theoretical approaches, as well as their arbitrator and conciliator. His depth of understanding of both approaches (he has "been there", after all) coupled with his clear writing style allows him to gently guide the reader through this fascinating and important area, engaging us with a seemingly whimsical but ultimately profound exercise in weighing elephants!

Even if you only buy one book on sample surveys, buy this one.

This wonderful book is posed to become THE reference on the subject of stochastic orders. It replaces a good but a bit outdated book by Shaked and Shanthikumar and even more outdated book by Stoyan. A great improvement over these two books is a large section devoted to applications of stochastic orders to a variety of practical models. I also liked concise diagrams showing relations among various orders as well as examples of distributions that can be ordered. This is definitely the most comprehensive reference that exists today.

"Typically a reader obtains their first impression of a new book by identifying the authors, the title and the blurb on the back, in that order, and to follow this algorithm would be to have the reader expect that the book is about statistical analysis in microarray experiments, alone. Even a cursory examination of the table of contents and skimming chapters and the color plates might perpetuate this view, but actually the book is much more than that"
From- Cancer Biology and Therapy

Reviewed by- Michael N. Liebman

This book provides a very clear treatment of model selection, the bootstrap and the computer-intensive philosophy. It places heavy emphasis on the bootstrap and cross validation. Other computer-intensive methods such as MCMC methods, multiple imputation and permutation methods are not treated. It represents development up to 1994 in a rapidly advancing area. There have been many additional theoretical and applied advances since then. It covers most of the topics of other competing books on bootstrap and resampling but is unique in its treatment of model selection and validation. It appears to require intermediate level knowledge of mathematics and freely uses matrix algebra. There are many good references and examples. The author uses FORTRAN to implement the methods and provides code. This may be a bit old fashioned.

In the course of my career as a research statistician particularly during my graduate school Ph.D. research I considered a number of multivariate or time series models that had specific marginal distributions. These models had some unpleasant properties (particularly the ones with exponential marginal distributions). Arnold, Castillo and Sarabia show here that the situation is much better when dealing with conditional specifications rather than marginal specifications. They provide an up-to-date treatment on the vast and scattered theory of conditional specification and characterization of multivariate distributions.

These results appear to have a variety of applications. It also appears to be a fertile area for interesting new research.

Chapter two on near and exact compatibility helps with problems in Bayesian solicitation of consistent prior distributions. There also appears to be many applications to bivariate and multivariate extremes, survival models and simultaneous equation models.

A classic, will survive through the ages as long as Real Analysis and Probability are studied by students the world over. You need it to get a fundamental grounding in Probability Theory