Multivariate Nonparametric Methods with R

Offers a fresh, fairly efficient, and robust alternative to analyzing multivariate data. This monograph provides an overview of the theory of multivariate nonparametric methods based on spatial signs and ranks. It uses marginal signs and ranks and different type of L1 norm.

---

This book introduces a new way to analyze multivariate data. The analysis of data based on multivariate spatial signs and ranks proceeds very much as does a tra- tional multivariate analysis relying on the assumption of multivariate normality: the L norm is just replaced by different L norms, observation vectors are replaced by 2 1 their(standardizedandcentered)spatial signsandranks,andso on.Themethodsare fairly ef?cient and robust, and no moment assumptions are needed. A uni?ed t- ory starting with the simple one-sample location problem and proceeding through the several-sample location problems to the general multivariate linear regression model and ?nally to the analysis of cluster-dependent data is presented. The material is divided into 14 chapters. Chapter 1 serves as a short introd- tion to the general ideas and strategies followed in the book. Chapter 2 introduces and discusses different types of parametric, nonparametric,and semiparametric s- tistical models used to analyze the multivariate data. Chapter 3 provides general descriptive tools to describe the properties of multivariate distributions and mul- variate datasets.
Multivariate location and scatter functionals and statistics and their use is described in detail. Chapter 4 introduces the concepts of multivariate spatial sign, signed-rank,andrank,and shows their connectionto certain L objectivefunc- 1 tions. Also sign and rank covariance matrices are discussed carefully. The ?rst four chapters thus provide the necessary tools to understand the remaining part of the book.