Measures and probabilities (Q1921004)

The author says: ``This book is intended to be an introductory, yet sophisticated, treatment of measure theory''. It has four parts: I. Integration relative to Daniell measures; II. Operations on measures defined on semigroups; III. Convergence of random variables; Conditional expectation; IV. Operations on Radon measures. Part I gives a version of Daniell's extension procedure as a common background to the integration theory based on set functions (Parts II and III) and Radon measures in Part IV which follows Bourbaki. The book gives sufficient details for an ambitious student and treats many specialized topics which makes it a useful reference. The total material is more than could be covered in one year. The exercises contain many concrete results of classical analysis (like proofs of the Stirling's formula and Hardy-Ramanujan theorem, Cauchy determinant). Here are some other special topics: vector valued integration, characterisation of Riemann integrable functions, Birkhoff's ergodic theorem, polar coordinates, Stieltjes integrals, order statistics (including central limit theorem for the empirical mean), Conditional densities, Haar measures, in particular on \(SO((n+1), \mathbb{R})\), Gelfand pairs. Here are some important topics the book does not include: Existence of conditional laws on Polish spaces and its connection to desintegration of measures, Kolmogorov's construction of projective limits of measures, liftings, classification of measurable material spaces. Altogether the book is a good source of interesting concrete examples and a reliable exposition of the general theory, trying to find a compromise between Bourbaki's view of Radon measures and basic needs of probability.