Multivariate Statistics: Old School is amathematical and methodological introduction to multivariate statistical analysis. It presents the basic mathematical grounding that graduate statistics students need for future research, andimportant multivariate techniques useful to statisticians in general. The material provides support forfurther study in big data and machine learning. Topics include The multivariate normal and Wishart distributions Linear models, including multivariate regression and analysis of variance, andboth-sides models (GMANOVA, repeated measures, growth curves) Linear algebra useful for multivariate statistics Covariance structures, including principal components, factor analysis, independence and conditional independence, and symmetry models Classification (linear and quadratic discrimination, trees, logistic regression) Clustering (K-means, model-based, hierarchical) Other techniques, including biplots, canonical correlations, and multidimensional scaling Most of the analyses in the book use the statistical computing environment R, for which there is an available package (msos)of multivariate routines and data sets. This text was developed over many years by the author, John Marden, while teaching in the Department of Statistics, University of Illinoisat Urbana-Champaign.

The authors have cleverly used exercises and their solutions to explore the concepts of multivariate data analysis. Broken down into three sections, this book has been structured to allow students in economics and finance to work their way through a well formulated exploration of this core topic. The first part of this book is devoted to graphical techniques. The second deals with multivariate random variables and presents the derivation of estimators and tests for various practical situations. The final section contains a wide variety of exercises in applied multivariate data analysis.

The majority of data sets collected by researchers in all disciplines are multivariate, meaning that several measurements, observations, or recordings are taken on each of the units in the data set. These units might be human subjects, archaeological artifacts, countries, or a vast variety of other things. In a few cases, it may be sensible to isolate each variable and study it separately, but in most instances all the variables need to be examined simultaneously in order to fully grasp the structure and key features of the data. For this purpose, one or another method of multivariate analysis might be helpful, and it is with such methods that this book is largely concerned. Multivariate analysis includes methods both for describing and exploring such data and for making formal inferences about them. The aim of all the techniques is, in general sense, to display or extract the signal in the data in the presence of noise and to find out what the data show us in the midst of their apparent chaos. An Introduction to Applied Multivariate Analysis with R explores the correct application of these methods so as to extract as much information as possible from the data at hand, particularly as some type of graphical representation, via the R software. Throughout the book, the authors give many examples of R code used to apply the multivariate techniques to multivariate data.

Mathematical Statistics: Old School covers three main areas: The mathematics needed as a basisfor work in statistics; the mathematical methods for carrying out statistical inference; and the theoretical approaches for analyzing the efficacy of various procedures. The author, John Marden,developed this material over the last thirty years teaching various configurations of mathematical statisticsand decision theory courses. It is intended as a graduate-level text. Topics include distribution theory, asymptotic convergence, frequentist and Bayesian inference (estimation, hypothesis testing, confidence intervals, model selection), exponential families, linear regression, likelihood methods, bootstrap and randomization methods, and statistical decision theory.

A traditional approach to developing multivariate statistical theory is algebraic. Sets of observations are represented by matrices, linear combinations are formed from these matrices by multiplying them by coefficient matrices, and useful statistics are found by imposing various criteria of optimization on these combinations. Matrix algebra is the vehicle for these calculations. A second approach is computational. Since many users find that they do not need to know the mathematical basis of the techniques as long as they have a way to transform data into results, the computation can be done by a package of computer programs that somebody else has written. An approach from this perspective emphasizes how the computer packages are used, and is usually coupled with rules that allow one to extract the most important numbers from the output and interpret them. Useful as both approaches are--particularly when combined--they can overlook an important aspect of multivariate analysis. To apply it correctly, one needs a way to conceptualize the multivariate relationships that exist among variables. This book is designed to help the reader develop a way of thinking about multivariate statistics, as well as to understand in a broader and more intuitive sense what the procedures do and how their results are interpreted. Presenting important procedures of multivariate statistical theory geometrically, the author hopes that this emphasis on the geometry will give the reader a coherent picture into which all the multivariate techniques fit.

This is the first book on multivariate analysis to look at large data sets which describes the state of the art in analyzing such data. Material such as database management systems is included that has never appeared in statistics books before.