Re: [Fwd: table model updates [corrected]]

Thanks Geoffrey,

That's really a very similar case to the one I'm asking about.
At first I thought the considertaion of X1min, X2max, .. etc. But then
came this case.

I think there is a complexity of constructing an N-dimensional surface
like a triangle in your 2D example below to encompass all points and
beyond which points should be extrapolated.

The first simple definition above seems computationally efficient,
That's all points beyond the 2D square should be extrapolated & any
point inside should be interpolated. But a true problem is there if
actual data occupy a small area of the sqaure.

Now, I don't know what should be done about this issue and that's why I
was asking.

Thanks,
Tarek

Geoffrey.Coram wrote:

>Tarek -
>Martin should really address this, as the original author,
>but he's out of the office for a while.
>
>I guess it's obvious that a point (x1,x2,x3, ...) will need
>to be extrapolated if x1 > x1max or x1 < x1min (or x2 > x2max,
>etc.)
>
>But a more interesting case (and probably what you were
>asking about) would be something like the following.
>Suppose we have data
>
># example.tbl
># x y f
> 0 0 0
> 1 0 1
> 1 1 2
> 2 0 4
> 2 1 5
> 2 2 8
>
>and the input point is (0.5,1.5) -- namely above the line x=y,
>whereas all the data is below that line. In this case, we have
>data points on both sides of the input point in both dimensions
>considered separately, but yet the point is "outside" the domain
>of data.
>
>-Geoffrey
>
>
>
>
>
>
>>Please I need to ask about extrapolation of data for more than 1D
>>case.
>>I think that for 1D case it's simple that if (Xin>Xmax) or
>>(Xin<Xmin) then Xin should be treated as a point needing
>>extrapolation.
>>But, for higher dimensions & given that the
>>data is scattered, how mathematically is a point defined as
>>needing extrapolation?
>>
>>Thanks,
>>Best Regards,
>>Tarek
>>
>>
Received on Mon Aug 9 04:30:00 2004