>Hi all VFP gurus,
>
>Has anyone a clue on the simplest way in calculate "least square" fitting from with VFP. The intention is pretty simple. A reasonable set of numeric (x,y) points available within VFP. We need the best fits for :
>- a straight line,
>- a logarithmic curve,
>- an exponential curve.
>
>As you imagine we do not want to use any VB stuff/dependancy. and are looking for the math formula not a visual representation. It could be straight VFP algorithms or dll-based library at a reasonable fee.
>
>Any clue or any pointer welcome!
>
>François
___
François,

Any of these helpful ?

straight line : http://www.bsu.edu/web/jkshim/mathandstat/lsm/leastsquare.htm under 3. Least Square Line which tells you how to compute a and b

also (12) (13) (14) (15) http://mathworld.wolfram.com/LeastSquaresFitting.html

logarithmic : http://mathworld.wolfram.com/LeastSquaresFittingLogarithmic.html

exponential: http://mathworld.wolfram.com/LeastSquaresFittingExponential.html