[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[EquisMetaStock Group] Re: Hotelling Transform and other matters



PureBytes Links

Trading Reference Links

--- In equismetastock@xxxxxxxxxxxxxxx, "formulaprimer" 
<formulaprimer@xxxx> wrote:
> I wanted to use it like the fisher transform but from what I 
> understand it is better used for smaller sample sets. Maybe using 
> multiple diminsions or matices it can simulate genetic algorism.
> Like parenting.  To difference parents combine to generate multiple
> outcomes, but only one fits the data the best or maybe forecast 
> method. Maybe you can't even use the Hotelling Transform for 
trading 
> <SNIP>

The "Hotelling Transform" is a tiny corner of a vastly "larger" 
universe of techniques. One big "chunk" is 'Orthogonal Functions'. 
The variety of orthogonal functions (i.e. basis functions)is 
defacto 'uncountable'. VERY many (with appropriate "twiddling") can 
be applied to time series analysis. The "easiest" forms to apply are 
orthogonal poynomials - Legendre, Tchebychev, Hermite, Meixner, 
Gegenbauer, etc. ad infinitum. The trigonometric varieties like DCT, 
Fourier, etc. are only different in their basis functions. The down 
side to the 'trig' type is the basis functions do not have 'compact 
support' (unlike wavelets and all their kin) while the series or 
signals they are applied to are of finite extent.
Regardless, actually CALCULATING the coefficients for an orthogonal 
decomposition of a 'time series' leads inevitably to Singular Value 
Decomposition (SVD) of Covar matrices (and/or other structures!). SVD 
BY ITSELF can be used (powerfully, usefully, and sucessfully) 
towards 'decomposition', 'noise reduction', 'smoothing' and all the 
other goals that simple garden variety "tools" like moving 
averages, "oscillators", etc. are pretenders to. I might add 
that "properly" decomposed time series (financial or otherwise) can 
produce particularly eye opening inputs for Neural Network prediction 
models and are very useful for PLS-2 (Partial Least Squares 'Two') 
regression models. It is truly endless. If anyone is interested, I 
can recommend (perhaps send) a variety of references and resources on 
these and related matters. I have a write-up overview on how a great 
deal of this 'stuff' can be applied in practice.
Finally, I have several/many of the techniques converted into 
Metastock DLLs. Out of respect to the members of this forum I have 
not "advertised" these things. If one has persisted in reading this 
far and has interest contact me via e-mail and I can offer more info 
on what I am developing, thinking about, etc.. I am most definitely 
NOT in the trading software business, nor do I intend to be. 
Regards,
WACG






------------------------ Yahoo! Groups Sponsor --------------------~--> 
Try Online Currency Trading with GFT. Free 50K Demo. Trade 
24 Hours. Commission-Free. 
http://us.click.yahoo.com/DldnlA/9M2KAA/U1CZAA/BefplB/TM
--------------------------------------------------------------------~-> 

 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/equismetastock/

<*> To unsubscribe from this group, send an email to:
    equismetastock-unsubscribe@xxxxxxxxxxxxxxx

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/