Customer Reviews

Statistical Analysis of Spherical Data

5 star:		(2)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

I you are just looking for a cookbook with ready to use recipes for the statistical analysis of directional data - this book is for you. It gives necessary rudiments of the relevant theory, but doesn't dive deeply into it. What it presents is algorithm-type descriptions covering various statistical test and parameter estimation techniques applied to directions (unit vectors or points on the sphere) and undirected data (axes). The use of these techniques is illustrated are with abundant and well described real life examples. The range of applications is extensive and covers virtually every single situation that could arise in routine work of a paleomagnetist, structural geologist, or any other Earth scientist using directional data. (Being an Earth scientist myself I cannot really testify how well it covers the aspect relevant to other fields, but I believe it is quite comprehensive.) These include

-- Simple exploratory and graphical techniques aimed to help in identifying an appropriate statistical model for the description of a specific sample of spherical data

-- Formal tests to see whether the data conform to a distribution of specific type (e.g., uniform, Fisher, Watson, Kent, Wood distributions)

-- General tests for uniformity and rotational symmetry

-- Simple non-parametric estimation of sample measures (e.g., mean, median, their confidence regions, etc.)

-- Parameter estimation for parametric models

-- Testing whether the sample could be drawn from a distribution with a specific properties (e.g, known mean, know dispersion)

-- Identifying outliers and formal tests for discordancy

-- Comparing two (or more) samples -- whether they could be drawn form the distribution, from distinct distributions with a common mean, common median, common concentration parameter, etc.

-- Estimation of common measures from two or more samples

-- Rudiments of correlation and regression analysis

-- Rudiments of bootstrap and other re-sampling techniques

What the book is missing entirely is the statistical treatment of data in plate tectonic reconstruction (estimation of tectonic rotations, their uncertainties, quality of fit etc.) This is not surprising because the book was written in 1987, when the relevant theory is only began to develop. For a more mathematically minded and those who are interested in the theory that lies beneath the techniques described in this book, I would recommend the "Directional Statistics" by Mardia and Jupp (1999, Whiley), which is also quite a pleasant read.

This book is a very good book which treated spherical data comprehensively.
Although written very carefully, if a fundamental description is added to a slight degree, the buyer of this book will increase more.
Thank you for a useful book.

The best to dream because it change my lif