home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math.stat
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!wupost!darwin.sura.net!Sirius.dfn.de!gwdu03.gwdg.de!ibm.gwdg.de!ALEWAND
- From: ALEWAND@ibm.gwdg.de
- Subject: Asymptotically chi-squared?
- Message-ID: <1682BDBFE.ALEWAND@ibm.gwdg.de>
- Sender: news@gwdu03.gwdg.de (USENET News System)
- Nntp-Posting-Host: ibm.gwdg.de
- Organization: GWDG, Goettingen
- Date: Tue, 21 Jul 92 15:38:38 MSZ
- Lines: 24
-
-
- Let X1,X2,.. be a sequence of k-dimensional random vectors (k>1),
- t1,t2,... a sequence of real vectors with lim tn=0, and
- sqrt(n) (Xn-tn)
- converges in distribution to the multinormal distribution N(0,SIGMA).
-
- SIGMA is idempotent with rank(SIGMA)=s and 1<=s<=k-1.
-
- My conjecture is:
-
- n (Transpose(Xn) Xn -Transpose(tn) tn) is asymptotically chi-squared
- with s degrees of freedom if and only if
-
- Lim n Transpose(tn) tn =0 (n to infinity).
-
- Is anybody able to prove (or disprove) this conjecture?
-
- Thanks
-
- Achim Lewandowski
- Institut fuer Statistik und Oekonometrie
- 3400 Goettingen
- Germany
- email: alewand@ibm.gwdg.de
-