Re: [boost] [ANN] Boost.Time_series: feedback requested

From: Jeff Garland (jeff_at_[hidden])
Date: 2006-12-12 01:00:14

Eric Niebler wrote:
> Jeff Garland wrote:
 >
>> In particular, I'm sorta surprised there isn't a moving average algorithm as 
>> that's time series algorithm with a really wide applicability.  Classic 
>> example -- calculate a 10 day rolling average of the closing price of some 
>> stock.  Or calculate a 20 minute rolling average of the trading prices or 
>> quotes.  And, so on...there are all sorts of other applications for rolling 
>> averages.
> 
> 
> Yep, that'd be a nice one to have. As I say in the description, the 
> library is light on algorithms at the moment; the key is the 
> infrastructure that makes the algorithms possible.

Yep...just a suggestion for future expansion :)

> Of
>> course, what I'm thinking is for the stock price rolling average example above 
>> you might want to use gregorian::date or posix_time::ptime to represent point 
>> of measurement of the stock price.  I wasn't finding the requirements for the 
>> offset_type...which is what I think would need to be replaced.
> 
> 
> Right, I should document the requirements on the offset_type. The code 
> has been tested with offset_types of int and double. I think the right 
> concept for offset_type is LessThanComparible.

Well, here's the thing.  In date-time the 'duration_type' and the 
'point_in_time_type' are different.  For example, with dates you have 'date' 
as the 'point_in_time' which is conceptually dimensionless and the 
duration_type 'day'.  If you subtract 2 points in time you get a duration. So 
it's possible there would need to be a different type in the mix to make 
things work.  Anyway, this will be a good experiment at some point. I'd do it 
myself if I only had more time.... ;-)

>> I also find the Discretization intervals interesting. Is that a standard way 
>> of handling the breakdown for statistical purposes?  The thing being that only 
>> days and weeks are actually fixed lengths in the *messy world*...months, 
>> quarters, etc are all different depending leap years and such.
>  
> The Discretization template parameter is mostly there just for type 
> checking your operations. You don't want to be adding a daily series 
> with a yearly one, because that doesn't make sense. The fact that some 
> years are leap years doesn't matter for the purpose of type checking.

Got it.

Jeff