7

Some friends and I (statistics grad students) would like to work through a good regression book. We're looking for

  • single and multiple linear regression
  • logistic regression
  • model selection
  • diagnostics
  • splines
  • penalized methods
  • classifiers, trees

Can anyone recommend a book (or books) that cover these topics?

Stephan Kolassa
  • 95,027
  • 13
  • 197
  • 357
Patrick McCarthy
  • 545
  • 4
  • 12
  • 2
    It doesn't appear that you're actually asking for a *regression* text, per se. But, the "standard" beginning graduate text that is going to cover all of those topics is [ESL](http://www-stat.stanford.edu/~tibs/ElemStatLearn/). Of course, covering so many topics necessarily means that either (a) the book is huge or (b) there is some lack of depth for each topic. ESL takes the latter approach (for the most part). – cardinal Jan 25 '13 at 00:34
  • 2
    Maybe some econometric texts? Graduate level are Greene "econometric analysis", Hayashi "econometrics" and very good for time series is Hamilton "time series analysis". They are quite complete though Greene and Hayashi more or less limit themselves to discrete time. – IMA Jan 25 '13 at 09:38
  • 2
    Frank Harrell's *Regression Modeling Strategies" might be a good choice for an applied perspective. – Charlie Jan 25 '13 at 15:48
  • @IMA Hayashi is imo the best graduate econometrics book in existence. Good thinking. – JohnK Jun 06 '14 at 14:33
  • Here's a similar thread on Wakefield's book. It seems like you have plenty of options, but that's another good one. https://stats.stackexchange.com/questions/86057/bayesian-and-frequentist-regression-methods-by-jon-wakefield-a-good-introduct – eric_kernfeld Mar 07 '21 at 17:24

1 Answers1

3

I believe Graham Cookson's answer to a similar question would be of assistance. Basically, he recommends Gelman and Hill's Data Analysis Using Regression and Multilevel/Hierarchical Models. According to Mr. Cookson, the book "covers basic regression, multilevel regression, and Bayesian methods in a clear and intuitive way" and "would be good for any scientist with a basic background in statistics".

Community
  • 1
Waldir Leoncio
  • 2,137
  • 6
  • 28
  • 42
  • Well, the OP asked for a graduate level book. Maybe Seber's book will do the job. http://www.amazon.com/Linear-Regression-Analysis-2nd-ed/dp/8126547286/ref=sr_1_1?s=books&ie=UTF8&qid=1455573774&sr=1-1&keywords=seber++linear I read in its time the first edition, and it is good. Do not know the second edition, though. – kjetil b halvorsen Feb 15 '16 at 22:04