home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math.stat:2630 sci.math.num-analysis:3618
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!cs.utexas.edu!sdd.hp.com!spool.mu.edu!agate!nima.berkeley.edu!shein
- From: shein@nima.berkeley.edu (Soren Hein)
- Newsgroups: sci.math.stat,sci.math.num-analysis
- Subject: Double decorrelation
- Date: 17 Dec 1992 15:54:24 GMT
- Organization: U.C. Berkeley -- ERL
- Lines: 20
- Distribution: world
- Message-ID: <1gq7rgINNcqk@agate.berkeley.edu>
- NNTP-Posting-Host: nima.berkeley.edu
-
- Here is one of those problems that appears to be simple, but
- apparently isn't:
-
- Given a matrix X (M by N) of stochastic variables, find two matrices
- A (M by M) and B (N by N) such that for the transformed matrix
-
- Y = A X B,
-
- we have
-
- E[Y Y^T] = I_M
- E[Y^T Y] = I_N
-
- where E[.] denotes expectation, ^T denotes transposition, and I_M and
- I_N are identity matrices of size M by M and N by N, respectively.
-
- /Soren
- shein@robotics.berkeley.edu
-
- (Needless to say, this is not a homework problem etc.)
-