Circular statistics in R (Q2844503)

It is well known that circular (directional) statistics is a branch of the discipline of statistics that deals with data for which the unit circle, torus, cylinder, sphere, hypersphere, etc., are natural supports. Circular statistics are useful when trying to analyze data that are taken from a distribution that circles back on itself. Thus the sample space is typically a circle or a sphere, so that standard methods for analyzing univariate or multivariate measurement data cannot be used. Special circular (directional) methods are required which take into account the structure of these sample spaces. This book gives a simple, systematic and unified treatment of the methodology of circular statistics. The authors give the theory underlying each technique, and illustrate applications by working through examples and using the R statistical package throughout. They focus on model fitting and also on hypotheses testing related with. A good range of models were presented by the authors. Computer-intensive resampling methods are explained. This book is suitable for graduate students, researchers and professionals across different fields of knowledgement, especially, biologists, sociologists, astronomers, physicists, geologists and architects. In my opinion, this book brings together simplicity and good organization. This properties make this book very useful for all levels of readers because it does not require a strong scientific background in order to understand it.NEWLINENEWLINEAlthough I like this book, I look forward to see other books dealing with the same subject but this contains advanced methods, such as Bayesian, multivariate analysis, nonparametric estimation and other methods.NEWLINENEWLINEThe book contains eight chapters. In the first one the authors present the R environment, circular statistics in other software environments, the aims of the book and the book structure and use. Chapter two deals with graphical representations of circular data since it contains raw plots of circular data, rose diagrams, kernel density estimates and linear histograms. Summery statistics of circular data is the subject of chapter three since this chapter contains sample trigonometric moments, measures of location (sample mean direction and sample median direction), measures of concentration and dispersion, measures of skewness and kurtosis and some other related subjects. In Chapter four the authors presented some models for circular random variables based on distribution theory. Basic inference for a single sample was the subject of chapter five. This chapter contains testing for uniformity, testing for reflective symmetry and inferences for key circular summaries.NEWLINENEWLINEChapter six contains fitting three distributions for a single sample, these distributions were the von Mises, Jones-Pewsey and inverse Batschelet. In chapter seven, the subject of comparing two or more samples of circular data is treated since it contains exploring graphical comparisons of samples, tests for a common mean direction, tests for a common median direction, tests for a common concentration, tests for a common distribution and Moore's test for paired circular data. Correlation and regression is the subject of chapter eight. This chapter contains in section one linear, circular associations (two of correlation coefficients, the first one due to Johnson, Wehrly and Mardia and the second one due to Mardia alone and is denoted by Mardia's rank correlation coefficient). Section two contains circular-circular associations (three of correlation coefficients, two of them due to Fisher and Lee and the other due to Jammalamadaka and Sarma), and also Rothman's test for independence. The subject of section four was the cosine regression model. Sections (8.5) and (8.6) deal with linear and circular regressors, respectively, for a circular response . Multivariate regression with circular regressors is the subject of section (8.7).