- #1
- 996
- 5
Anybody know the math/theory behind linear least squares where the curve is forced to go through the first and last data points?I'm specifically dealing with cubic polynomials.
In standard linear least squares formulation (i.e. ATAc = ATy) the curve doesn't, in general, go through any of the data points. I've actually managed to come up with a formulation based on Bezier math that allows standard least squares treatment to be used, but the development is problem dependent and rather long and tedious.
Just wondering if there is an easiser way. Ideas?
In standard linear least squares formulation (i.e. ATAc = ATy) the curve doesn't, in general, go through any of the data points. I've actually managed to come up with a formulation based on Bezier math that allows standard least squares treatment to be used, but the development is problem dependent and rather long and tedious.
Just wondering if there is an easiser way. Ideas?
Last edited: