Customer Reviews

Discriminant Analysis (Quantitative Applications in the Social Sciences)

5 star:		(2)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

This is an excellent in depth introduction to Discriminant Analysis. This is the second little green book of the Sage University Paper I have studied. The first one was on Logistic Regression. And, the mileage you will get out of these green books will vary based on the complexity of the subject.

Frankly, by comparison Logistic Regression was a cruise. After reading the green book I was able to build an Excel based model longhand replicating most of the methodologies described in the book.

Discriminant Analysis is different. Relative to Logistic Regression it is a real piece of work. At first, I thought this green book was not as well written as the one on Logistic Regression. But, that is not the case. It is just that Discriminant Analysis is that much more complex. It is a bit like comparing the clarity of a physics textbook vs. a quantum physics one. The quantum physics one will appear much more obscure. And, it is not because of the writing style, but because of the subject complexity.

With Discriminant Analysis, you just have to accept a longer more intense learning curve than with Logistic Regression. Given that, this book is excellent. And, it gives you a pretty good understanding of what it is all about. It is a bit like learning how to drive. You can do it without understanding all the mechanical complexities of a combustion engine. This green book will give you a pretty good understanding on how to "drive" and interpret Discriminant Analysis. That probably meets the goals of 90% of the people who approach this Byzantine subject. If you are among the extreme mathematicians who want to be able to derive all the underlying algorithms, you will need more advanced books. And, the green book gives you ample references if you care to do that.

Like all of the Sage books I've read, the writing in this book very much gets to the point: the entire text is less than 70 pages long! Mainly, this book covers: canonical discriminant analysis and linear and quadratic discriminant classifiers, though a number of ancillary topics are also covered, such as variable selection and violations of assumptions. A capable programmer should be able to implement the described procedures without difficulty.

I would have preferred the inclusion of more worked examples and computer source code, otherwise this is an excellent book.