RE: cov. matrix of a stabilized wlsq fit

Article View: bit.listserv.stat-l

Article #3535

Back to bit.listserv.stat-l Previous Next

#3535

From: pgill
Date: Wed, 12 Feb 1992 15:32

12 lines
948 bytes

        In general,let your estimator be

        x = Gy , where matrix G does not involve y.  Then
the covariance matrix of your estimator is
        Cov(x) = GCG'.

In your second case, the covariance matrix is

        Cov(x) = inv(A'*W*A+alpha*E)*A'*inv(A'*W*A+alpha*E)'

P. S. Gill
Memorial Univ of Newfoundland, Canada

Technical Details

Message-ID: <009560AE.C1653860.31248@Leif.ucs.mun.ca>

Path: arc1-ca.pugleaf.net!sparky!uunet!paladin.american.edu!auvm!KEAN.UCS.MUN.CA!PGILL

Show Full Headers

Complete Article Headers

X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: f5efb,445f9541b3834349
X-Google-Attributes: gidf5efb,public
X-Google-ArrivalTime: 1992-02-12 06:51:19 PST
Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
Path: arc1-ca.pugleaf.net!sparky!uunet!paladin.american.edu!auvm!KEAN.UCS.MUN.CA!PGILL
Message-ID: <009560AE.C1653860.31248@Leif.ucs.mun.ca>
Newsgroups: bit.listserv.stat-l
Date:         Wed, 12 Feb 1992 10:32:00 -0500
Reply-To:     pgill@kean.ucs.mun.ca
Sender:       "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
From:         pgill@KEAN.UCS.MUN.CA
Subject:      RE: cov. matrix of a stabilized wlsq fit