Read more
This book provides an introduction to various aspects of Algebraic Statistics with the principal aim of supporting Master's and PhD students who wish to explore the algebraic point of view regarding recent developments in Statistics. The focus is on the background needed to explore the connections among discrete random variables. The main objects that encode these relations are multilinear matrices, i.e., tensors. The book aims to settle the basis of the correspondence between properties of tensors and their translation in Algebraic Geometry. It is divided into three parts, on Algebraic Statistics, Multilinear Algebra, and Algebraic Geometry. The primary purpose is to describe a bridge between the three theories, so that results and problems in one theory find a natural translation to the others. This task requires, from the statistical point of view, a rather unusual, but algebraically natural, presentation of random variables and their main classical features. The third part of the book can be considered as a short, almost self-contained, introduction to the basic concepts of algebraic varieties, which are part of the fundamental background for all who work in Algebraic Statistics.
List of contents
PART I: Algebraic Statistics.- 1 Systems of Random Variables and Distributions.- 2 Basic Statistics.- 3 Statistical models.- 4 Complex projective algebraic statistics.- 5 Conditional independence.- PART II: Multilinear Algebra.- 6 Tensors.- 7 Symmetric tensors.- 8 Marginalisation and attenings.- PART III: Commutative Algebra and Algebraic Geometry.- 9 Elements of Projective Algebraic Geometry.- 10 Projective maps and the Chow's Theorem.- 11 Dimension Theory.- 12 Secant varieties.- 13 Groebner bases.
About the author
Prof. Luca Chiantini is Full Professor of Geometry at the University of Siena (Italy). His research interests focus mainly on Algebraic Geometry and Multilinear Algebra, and include the theory of vector bundles on varieties and the study of secant spaces, which are the geometric counterpart of the theory of tensor ranks. In particular, he recently studied the relations between Multilinear Algebra and the theory of finite sets in projective spaces.
Prof. Cristiano Bocci is Assistant Professor of Geometry at the University of Siena (Italy). His research concerns Algebraic Geometry, Commutative Algebra, and their applications. In particular, his current interests are focused on symbolic powers of ideals, Hadamard product of varieties, and the study of secant spaces. He also works in two interdisciplinary teams in the fields of Electronic Measurements and Sound Synthesis.
Summary
This book provides an introduction to various aspects of Algebraic Statistics with the principal aim of supporting Master’s and PhD students who wish to explore the algebraic point of view regarding recent developments in Statistics. The focus is on the background needed to explore the connections among discrete random variables. The main objects that encode these relations are multilinear matrices, i.e., tensors. The book aims to settle the basis of the correspondence between properties of tensors and their translation in Algebraic Geometry. It is divided into three parts, on Algebraic Statistics, Multilinear Algebra, and Algebraic Geometry. The primary purpose is to describe a bridge between the three theories, so that results and problems in one theory find a natural translation to the others. This task requires, from the statistical point of view, a rather unusual, but algebraically natural, presentation of random variables and their main classical features. The third part of the book can be considered as a short, almost self-contained, introduction to the basic concepts of algebraic varieties, which are part of the fundamental background for all who work in Algebraic Statistics.
Additional text
“The book under review contributes to the literature in algebraic statistics by highlighting the role of tensors, which are a vital tool in algebraic statistical theory and methods.” (Carlos Amendola, Mathematical Reviews, December, 2020)
Report
"The book under review contributes to the literature in algebraic statistics by highlighting the role of tensors, which are a vital tool in algebraic statistical theory and methods." (Carlos Amendola, Mathematical Reviews, December, 2020)
