| re: Formula required
On Thu, 27 Oct 2005 03:36:33 GMT, "Larry Linson" <bouncer@localhost.not> wrote:
[color=blue]
>"Tom van Stiphout" wrote
>[color=green]
> > That depends. I look at this as a Math
> > problem, with some X,Y values, and
> > some additional X-values where we want
> > to know the Y-value. Typically you'd want
> > to do some sort of interpolation. Perhaps
> > straight lines between the known points
> > (linear interpolation), perhaps a smooth
> > line (e.g. a cubic spline).[/color]
>
>But, without more information on what kind of data it is and/or how it is
>derived, we couldn't suggest what method would be appropriate. I remember
>taking a numerical analysis class in which, to point out the fallacy of
>assuming the missing values could be calculated with whatever method we were
>learning, the teacher gave a similar list with some missing values. It
>turned out that no one's suggested method would work... it was a list of
>room numbers in the building and the number of chairs/seats in the room.
>
>And, as the missing numbers were associated with things like broom closets,
>other janitorial rooms, and storage rooms (which contained neither seats nor
>chairs) and one storage room that was stacked, jammed-full, floor to ceiling
>with extra folding chairs (which contained all the extra chairs for the
>entire building)... there was no "method" that could be used to _calculate_
>them.
>
>If the ghanley's teacher gave this as an assignment, ghanley had best go ask
>for more information. If it is some real-world situation, then ghanley
>should carefully consider if it is a problem that lends itself to a
>calculated solution.
>
> Larry Linson
> Microsoft Access MVP
>[/color]
Larry is exactly right. Thers is no reasonable relationship between X and its
corresponding value of Y. A least spline that passes thru all the data points
looks weird. That said, it is possible to write a funftion generator that
doesn't look too bad and which will pass thru all of the given data points.
However, the calculated values of Y for values of X between the original data
points is pure fiction. I believe you have coppied the data incorrectly or
your instructor has thrown you a mickey.
75000 ~~ 806.2-----
33000 ~~ 592.0-----
8200 ~~ 382.7-----
345 ~~ 33.1-----
Chuck
--
| 
