Matrix Analysis for Statistics

Rating:
Summary: Use symbols in a confusing way
Review: I can't recommend this book.

One problem is that the author uses symbol in a confusing way: declare without define, define without explaination, especially in the examples related to linear regression. It seems he just pick up symbols at random whenever he needs one in the middle of a proof, make an equation out of it. It's your role to figure out from the equation what this symbol is and why the author introduce this symbol.
More than that symbols are not used consistently. Z1, for example, is used to represent column vectors, row vectors, orthornormal vectors and centeralized observation vectors in very close sections without explanation. A vector just defined as a column vector can be multiplied with another column vector. Seems the author doesn't pay much attentions to such "small" aspects and this makes the book really hard to read.

Another problem is you won't know what the author is trying to do only after the whole algebraic series is finished.The objective is always the last to come. Won't it be nice to just add a few words before each a few steps to say what you are planning to do?

The last complaint is that there's not a simple geometric graph in this book. How can you imagine a book on the subject of both statistics and matrix don't have a illustrative graph on it?