Tensor Methods In Statistics: Second Edition Av .Peter Mccullagh

A pioneering monograph on tensor methods applied to distributional problems arising in statistics, this work begins with the study of multivariate moments and cumulants. An invaluable reference for graduate students and professional statisticians, it requires some knowledge of linear algebra, eigenvalue decompositions, linear models, and, for later chapters, likelihood functions and likelihood ratio statistics. More than 200 exercises throughout the book are an integral part of the text. 1987 edition.