Order from a Major Online Bookstore
Ultimate Book Price Comparison Engine
Save Your Time And Money
| AnyBook4Less.com | Order from a Major Online Bookstore |
|
| Home | Store List | FAQ | Contact Us | | ||
| Ultimate Book Price Comparison Engine Save Your Time And Money |
||
 |
Title: Subset Selection in Regression, Second Editon by Alan J. Miller ISBN: 1-58488-171-2 Publisher: Chapman & Hall/CRC Pub. Date: 15 April, 2002 Format: Hardcover Volumes: 1 List Price(USD): $89.95 |
Average Customer Rating: 4.5 (2 reviews)
Rating: 5
Summary: This is a book you must own
Comment: There's a revolution of sorts in statistical modeling in recent years, first from recognizing the effects of model selection bias on inference and prediction, then from the impetus of boosting and then bagging from cs, statistical modeling is probably not the same since. The earlier version of book (1990) was the first book discussing why practioners involving modeling with many variables need to pay attention to some of the deep issues. The present updated edition incorporates some of the recent developments in last 10 years, noticeably, penalized likelihood type methods, Bayesian method, and bagging type procedure. I believe there will be more exciting developments in this area, including the Bayesian approach, which may offer the ultimate hope of providing valid inference and prediction procedure. In all, the second edition of this book provides a golden opportunity for researchers who haven't already own it to buy and to read and to enjoy it.
Rating: 4
Summary: Best Available But Difficult to Read
Comment: This book describes techniques for finding good predictors of some phenomenon (assuming linear dependence). For example, suppose you want to know what factor determine IQ. You do a survey that includes 500 factors such as "family income", "length of hair", "father's IQ", etc. Now determine which are the 20 best determinants of IQ! Thats impossible (practically speaking) because you'd have to do an exhaustive search of 500 choose 20 variables which would take to much computing power. The book describes known techniques which identify factors that are good predictors (but not necessarily the best).
In my searches, I found this book to be the best available on the subject. I found it difficult to read -- it is math heavy with emphasis on linear algebra.
The author is certainly one of the most authoritative subject matter experts. He supports free fortran source code which are available on his WEB site. He also maintains software (for MATHLAB I think) which use the techniques.
I would like the book to be targetted at three levels of audiences (I know - I want to world!). The book should include the math - for the theoretician. The book should be organized so persons not interested in the theory can figure out what the techiques do (at a high level). Finally, the book should identify which methods are supported by the major commerical statistical packages (e.g., SPSS, MATLAB, etc).
 |
Title: Model Selection and Multi-Model Inference by Kenneth P. Burnham, David Anderson ISBN: 0387953647 Publisher: Springer-Verlag Telos Pub. Date: 12 July, 2002 List Price(USD): $84.95 |
 |
Title: Bayesian Data Analysis, Second Edition by Andrew Gelman, John B. Carlin, Hal S. Stern, Donald B. Rubin ISBN: 158488388X Publisher: Chapman & Hall/CRC Pub. Date: 29 July, 2003 List Price(USD): $59.95 |
 |
Title: Applied Logistic Regression (Wiley Series in Probability and Statistics - Applied Probability and Statistics Section) by David W. Hosmer, Stanley Lemeshow ISBN: 0471356328 Publisher: Wiley-Interscience Pub. Date: 15 September, 2000 List Price(USD): $105.00 |
Thank you for visiting www.AnyBook4Less.com and enjoy your savings!
Copyright� 2001-2021 Send your comments
