home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!howland.reston.ans.net!europa.asd.contel.com!paladin.american.edu!auvm!CUA.BITNET!CHENX
- Original_To: BITNET%"sas-l@uga"
- Message-ID: <SAS-L%93010616584010@UGA.CC.UGA.EDU>
- Newsgroups: bit.listserv.sas-l
- Date: Wed, 6 Jan 1993 16:57:00 EST
- Reply-To: Chen Xi <CHENX@CUA.BITNET>
- Sender: "SAS(r) Discussion" <SAS-L@UGA.BITNET>
- From: Chen Xi <CHENX@CUA.BITNET>
- Subject: Re: ML vs OLS in regression
- Lines: 6
-
- At first, when the error is assumed to be normal, ML and OLS are
- equivalent. Without the normal assumption, they might be different.
- One of the important application of the ML estimator is the model
- selection with information theoretical criteria. When the evaluation of
- the model is based on the entropy, the maximum likelihood estimation
- is required.
-