In fn:= (L'delta)^2/(L' Sigma L) multiplying L by any constant does not change
anything. Thus we need to fix the scale at something to get a unqiue solution.
At 3:25 PM +0000 11/22/99, Brad wrote:
>Thanks again. It's beginning to make more sense. However, I had one final
>question:
>
>Why did you assume that L'SigmaL=1?
>
>Brad
>
>In article <v04220814b45e7d384528@[204.212.38.166]>,
> [EMAIL PROTECTED] (Jan de Leeuw) wrote:
> > Maximize l'delta delta'l on the condition that l'Sigma l=1. Use a
> > Lagrange multiplier
> > lambda to find delta delta'l = lambda Sigma l, which shows Sigma
>l proportional
> > to delta.
> >
> > At 2:47 AM +0000 11/22/99, Brad wrote:
> > >Thanks. You're right. The numerator should be (L'delta)^2.
>However, I'm still
> > > confused about the solution. How did you arrive at the solution?
> > >
> > >Thanks.
> > >brad
> > >
> > >In article <v04220800b45c9787eb74@[204.212.38.187]>,
> > > [EMAIL PROTECTED] (Jan de Leeuw) wrote:
> > > > This can't be right. Let L->0 then fn -> infty. Probably you want
> > > > sqrt(L'Sigma L) in the denom. In that case, the solution is simply
> > > > proportional to sigma^{-1}delta.
> > > >
> > > > At 4:42 PM +0000 11/20/99, Brad wrote:
> > > > >Hello.
> > > > >
> > > > >I'm trying to maximize (wrt L) the following function:
> > > > >
> > > > >fn:= (L'delta)/(L' Sigma L)
> > > > >
> > > > >where L = p x 1
> > > > > delta = p x 1
> > > > > Sigma = p x p
> > > > >
> > > > >I'd appreciate any help in the matter.
> > > > >
> > > > >Brad
> > > > >
> > > > >
> > > > >Sent via Deja.com http://www.deja.com/
> > > > >Before you buy.
> > > >
> > > > ===
> > > > Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
> > > > US mail: 8142 Math Sciences Bldg, Box 951554, Los Angeles,
>CA 90095-1554
> > > > phone (310)-825-9550; fax (310)-206-5658; email:
>[EMAIL PROTECTED]
> > > > http://www.stat.ucla.edu/~deleeuw and
>http://home1.gte.net/datamine/
> > > >
> > >======================================================================
> > >======
> > > > No matter where you go, there you are. --- Buckaroo Banzai
> > > >
> > >======================================================================
> > >======
> > > >
> > >
> > >
> > >Sent via Deja.com http://www.deja.com/
> > >Before you buy.
> >
> > ===
> > Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
> > US mail: 8142 Math Sciences Bldg, Box 951554, Los Angeles, CA 90095-1554
> > phone (310)-825-9550; fax (310)-206-5658; email: [EMAIL PROTECTED]
> > http://www.stat.ucla.edu/~deleeuw and http://home1.gte.net/datamine/
> >
>======================================================================
>======
> > No matter where you go, there you are. --- Buckaroo Banzai
> >
>======================================================================
>======
> >
>
>
>Sent via Deja.com http://www.deja.com/
>Before you buy.
===
Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
US mail: 8142 Math Sciences Bldg, Box 951554, Los Angeles, CA 90095-1554
phone (310)-825-9550; fax (310)-206-5658; email: [EMAIL PROTECTED]
http://www.stat.ucla.edu/~deleeuw and http://home1.gte.net/datamine/
============================================================================
No matter where you go, there you are. --- Buckaroo Banzai
============================================================================