No label defined (Q4338926)

Regression is probably the most used methodology in statistics; modern computer technology has brought this area on new levels, research on it grows exponentially fast. This book successfully attempts to give an overview of the state-of-the-art. In my opinion it is most helpful for researchers and practitioners who already know about regression analysis and want to be informed about latest developments. (From the references there are more than thirty later than 1992.) They will find an immense amount of material -- carefully analyzed and explained, with excellent statistical intuition, sometimes highly critical -- with discussions of advantages and disadvantages, hints on remedies and well chosen examples. A highlight is the last chapter on applications where for five real data sets almost all aspects of a thorough statistical analysis are discussed. Although there is a wealth of exercises, I think its usefulness as textbook is limited. Too much is referred to the literature, most of the concepts are explained only by example. Good knowledge of general statistical methods is indispensable. One who likes formal approaches will be disappointed, one who wishes to learn about good statistical reasoning and actual developments in an important branch of statistics will gain much from it.