[OT] What field (if any) of math am I looking for?



Dan Hatton wrote:
> On Mon, 26 May 2008, Richard Owlett wrote:
> 
> 
>>In an engineering group, there was a question of comparing the "shape"
>>of plots of observed data. The discussion got tangled in "where did data
>>come from" rather than whether or not one short sequence (tens of
>>points) in some sense had the same "shape" as another sequence.
>>
>>I attempted to recast the problem
>>
>>Given pairings of datasets ( ALPHA  BETA ), ( BETA GAMMA) etc
>>is there a measure ( OMEGA(x,y) ) such that
>>
>>OMEGA(ALPHA(i), ALPHA(j)) always greater than OMEGA(ALPHA(i), BETA(k))
> 
> 
> Substitute "always" for "with some probability that we can estimate",
> and you might be looking for "classifiers" or "model comparison".  In
> which case, there's some relevant material in
> 
> D. J. C. MacKay.  Information theory, inference, and learning
> algorithms.  Cambridge University Press, Cambridge, Sep. 2003.  URL:
> <http://www.inference.phy.cam.ac.uk/itprnn/book.pdf>;.
> 
> Pulling this back on topic, a quick scan of the Maxima manual suggests
> the lsquares and stats packages are relevant.
> 

Thank you, and to Stavros Macrakis who replied off list.
You both have have given me the search terms I needed to start digging.