Linear Statistical Inference and Its Application

Rating:
Summary: This is a classic by a maestro
Review: This book contains an encyclopaedic wealth of material on the mathematical theory of Statistics upto the beginning of the 70's, packed in its 600-odd pages. It contains material for several courses in Statistics at both undergraduate and graduate levels. In Chapter 1 it captures in a few pages the essence of matrix theory required for Statistics. The material on Linear Models and Least Squares in Chapter 4 is excellent. Chapter 8 captures succinctly in a few pages the entire theory of the multivariate normal and sampling distributions thereof using the characterisation arising from the Cramer--Wold Theorem. The other chapters too contain a phenomenal amount of material in a wide sweep of the theory of Statistics. However, it is not suitable as a text-book since the writing is terse; an instructor who already knows the subject will benefit from the book to prepare his material; students will find it difficult to read it on their own. Further, the proof-reading is somewhat shoddy and there are many typographical errors. Then it has some idiosyncracies like presenting the author's list of publications, impressive though it is. If you are a working statistician or a researcher or a teacher of Statistics, you have to have a copy of it on your desk within easy reach!