Introduction to Graphical Modelling (Springer Texts in Statistics) |
 | Author: David Edwards
Publisher: Springer
List Price: $64.95
Buy Used: $44.38
as of 3/17/2010 22:03 CDT details
You Save: $20.57 (32%)
New (2) Used (13) from $44.38
Seller: thebookgrove
Rating:  1 reviews
Sales Rank: 4478279
Media: Hardcover
Edition: 1
Pages: 274
Number Of Items: 1
Shipping Weight (lbs): 1.6
Dimensions (in): 9.5 x 7.3 x 1
ISBN: 0387944834
Dewey Decimal Number: 519.538
EAN: 9780387944838
ASIN: 0387944834
Publication Date: April 21, 1995
Availability: Usually ships in 1-2 business days
|
|
|
Also Available In:
|
|
Similar Items:
|
|
|
Editorial Reviews:
Product Description
Graphical modelling is a form of multivariate analysis that uses graphs to represent models. They enable concise representations of associational and casual relations between variables under study. This textbook provides an introduction to graphical models whose emphasis is on its applications and on the practicalities rather than a formal development. With the book comes a diskette containing a student version of MIM - a popular graphical modelling software package for the PC. Following an introductory chapter which sets the scene and describes some of the basic ideas of graphical modelling, subsequent chapters describe particular families of models including log-linear models, Gaussian models, and mixed models for discrete and continuous data. Further chapters cover hypothesis testing for mixed models and discuss issues of model selection and more advanced topics.
|
|
Customer Reviews:
directed graphs, path analysis and causality not the common statistical graphics February 17, 2008
Michael R. Chernick (Holland PA)
30 out of 30 found this review helpful
Because graphic methods are very popular in statistics, when you read the title you might think this is a book on the use of graphics in statistics. That is not what the book is about. The directed graph on the cover might be a hint for some.
The book deals with the theory of undirected and directed graphs which has applications to causal modeling in statistics and the development of expert systems (which Edwards claim are now more commonly referred to as probabilistic networks).
This subject is being made popular again based on the recent work of Edwards, Pearl, Rubin and a few others. The book incorporate the approach in many classical statistical problems. This is not commonly seen except in specialized texts on latent variable models.
Edwards discusses implementation of the methods with the freeware MIMS that is available in Denmark and on the web. The book is very well written and applications in MIMS are given throughout the text. Edwards also provides us with an excellent list of references (over 200 with many on causal modeling).
The software LISREL produced by researchers in the US at UCLA for latent variable and path analyses is only briefly mentioned on page 217. The lack of coverage of American and British publications on this topic is the only drawback I see.
|
|
