adaptive filters

"Dave" <dspguy2@netscape.net> wrote in message 
news:041fd32c-582e-48d6-b09e-644ce94b0817@kt16g2000pbb.googlegroups.com...
On Dec 5, 2:34 pm, "Vladimir Vassilevsky" <nos...@nowhere.com> wrote:
> "Dave" <dspg...@netscape.net> wrote"
> On Dec 5, 10:59 am, "Vladimir Vassilevsky" <nos...@nowhere.com> wrote:

Vlad - Grow up ! If you can't make a reasonable contribution then go
back and crawl under whatever rock you came from.

Dave 

On Dec 5, 2:56&#4294967295;pm, "Vladimir Vassilevsky" <nos...@nowhere.com> wrote:

Least mean squares is teh same as

On Thursday, December 6, 2012 4:59:47 AM UTC+13, Vladimir Vassilevsky wrote:
> <gyansorova@gmail.com> wrote:
> 
> 
> 
> 
> 
> >The LMS isn't a least squares fit -
> 
> 
> 
> Idiot.
> 
> LMS = least mean squares.
> 
> 
> 
> > it uses an instantaneous estimate
> 
> >of the gradient to try and find a minimum.
> 
> 
> 
> Using instantaneous gradient = SG method (stochastic gradient).
> 
> 
> 
> >actually LMS is minimising the mean-square error by gradient descent so is 
> 
> >a least-squares fit as well as the RLS.
> 
> 
> 
> If the statistics of signal is gaussian, then SG method converges to LMS 
> 
> solution.
> 
> 
> 
> >Difference is that RLS can be applied to pole-zero models whereas LMS is 
> 
> >all zero models (normally).
> 
> 
> 
> SG method could be used for pole-zero models as well. The computation of the 
> 
> gradient is somewhat more complicated then in all zero case, as it requires 
> 
> derivatives of denominator.
> 
> 
> 
> Vladimir Vassilevsky
> 
> DSP and Mixed Signal Consultant
> 
> www.abvolt.com

least mean squares is the same as least-squares, only it is defined for guassian processes. You are the idiot. The method of convergence may be different but the end result is the exact same. So instead of inverting matrices, recursive or otherwise, you minimise a gradient. The path may be different but the result is the same.


Hardy

gyansorova@gmail.com wrote:
> Least mean squares is teh same as
 
> On Thursday, December 6, 2012 4:59:47 AM UTC+13, Vladimir Vassilevsky wrote:

(snip)
>> LMS = least mean squares.

(big snip)
> least mean squares is the same as least-squares, only it is 
> defined for guassian processes. You are the idiot. The method 
> of convergence may be different but the end result is the 
> exact same. So instead of inverting matrices, recursive or 
> otherwise, you minimise a gradient. The path may be different 
> but the result is the same.

Many non-linear minimization problems can easily get stuck
in a local minimum instead of the desired global minimum.

In that case, the result isn't always the same, though the
desired result might be.

Also, even if the eventual minimum is the same, some algorithms
converge to it much faster than others.

-- glen

On Friday, December 7, 2012 9:24:36 AM UTC+13, glen herrmannsfeldt wrote:
> gyansorova@gmail.com wrote:
> 
> > Least mean squares is teh same as
> 
>  
> 
> > On Thursday, December 6, 2012 4:59:47 AM UTC+13, Vladimir Vassilevsky wrote:
> 
> 
> 
> (snip)
> 
> >> LMS = least mean squares.
> 
> 
> 
> (big snip)
> 
> > least mean squares is the same as least-squares, only it is 
> 
> > defined for guassian processes. You are the idiot. The method 
> 
> > of convergence may be different but the end result is the 
> 
> > exact same. So instead of inverting matrices, recursive or 
> 
> > otherwise, you minimise a gradient. The path may be different 
> 
> > but the result is the same.
> 
> 
> 
> Many non-linear minimization problems can easily get stuck
> 
> in a local minimum instead of the desired global minimum.
> 
> 
> 
> In that case, the result isn't always the same, though the
> 
> desired result might be.
> 
> 
> 
> Also, even if the eventual minimum is the same, some algorithms
> 
> converge to it much faster than others.
> 
> 
Rate of convergence is different - of course. there are trade-offs and many of them. For linear systems there are not any local minimums of course. we can always find problems if we dig deep, but the basic idea is the same, least squares. If he had said H infinity minimisation then I would have agreed, there you are minimising the maximum. Not only will the path be different but also the result.