0

I have a large number of x, y, z point values. I am wanting to know the equation that represents these values I have in hand. This equation I am looking for should be a fit that will represent these data points.

Which method(s) may I use in order to construct the equation that belongs to this specific data set? Is there a specific method to find the coefficients of x, y and z in an equation that fits a certain set of data?

Story Maths
  • 21
  • 1
  • 4
  • There could be uncountably many such equations. Normally you model it - ie assume a form for the equation with parameters to be determined. Then you fit an equation of that form by determining the parameters using the data. With enough parameters and complex enough equation form you may be able to fit every point perfectly, but usually that is not want you want. Usually you want to see if a simple form with few parameters fit "well enough" using some definition of good fit. This subject is too broad unless you narrow down your question further - say by seeking linear equations only. – Macavity Apr 12 '16 at 18:38
  • Thanks for your comment, I actually am looking for a "well enough" fit as you mentioned. If I could find the coefficients of x, y and z (that of the equation that fits the data set well enough) using the data I have, that would be great. I will add this information to the question to clarify the question. – Story Maths Apr 12 '16 at 18:42
  • If it's a linear equation you want, check http://mathforum.org/library/drmath/view/69103.html or http://stackoverflow.com/questions/24747643/3d-linear-regression – Macavity Apr 12 '16 at 18:47
  • If it's not linear, you may want to look into some kind of manifold discovery algorithm. – Brian Tung Apr 12 '16 at 20:13

0 Answers0