Article View: sci.math.stat
Article #605Re: robust location estimators
From: Jerry Dallal
Date: Thu, 01 Oct 1992 14:30
Date: Thu, 01 Oct 1992 14:30
18 lines
1405 bytes
1405 bytes
In article <44800@ucbvax.BERKELEY.EDU>, beran@wizzle.Berkeley.EDU (ibidem) writes: >> ... why robust location estimators have not become more popular. > > One answer might be: too many candidates, with too many differing > theoretical optimality properties. As it stands, the problem of > robust estimation is not reasonably well-defined. It will take a > new approach, I suspect, at some future date. And because of the lack of mass market software. Look how popular Cleveland's LOESS has become (in some circles) now that many stat packages will computer it. The same thing seems to be happening with Rousseeuw's PROGRESS. Even $10 calculators have sigma buttons. Haven't seen one with an M-estimator button, yet. If you want to make advance a new statistical technique, don't worry about writing theory articles for prestigious statistical publications. Just write some easy-to-use software and market it like mad.
Technical Details
Message-ID:
<1992Oct1.093017.409@hnrc.tufts.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!usc!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!bu.edu!news.tufts.edu!news.hnrc.tufts.edu!jerryReferences:
<19s96oINNdd9@agate.berkeley.edu> <44800@ucbvax.BERKELEY.EDU>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!usc!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!bu.edu!news.tufts.edu!news.hnrc.tufts.edu!jerry From: jerry@hnrc.tufts.edu (Jerry Dallal) Newsgroups: sci.math.stat Subject: Re: robust location estimators Message-ID: <1992Oct1.093017.409@hnrc.tufts.edu> Date: 1 Oct 92 14:30:17 GMT References: <19s96oINNdd9@agate.berkeley.edu> <44800@ucbvax.BERKELEY.EDU> Organization: USDA HNRC at Tufts University Lines: 18