2

My data points are (97.57,6.14), (90.54,7.03), (81.99,8.55), (71.47,10.52), (56.5,14.97) and (31.88,24.62). I'm trying to find the nonlinear equation that describes these points, but I'm having trouble. Can anyone come up with it.

john griffin
  • 21
  • http://www.wolframalpha.com/input/?i=fit+%2897.57%2C6.14%29%2C+%2890.54%2C7.03%29%2C+%2881.99%2C8.55%29%2C+%2871.47%2C10.52%29%2C+%2856.5%2C14.97%29%2C+%2831.88%2C24.62%29 – Amzoti Jul 14 '13 at 03:49
  • 3
    There's no such thing as "the" equation fitting the data. Lagrange interpolation will give you "a" function fitting the data; cubic splines will give you another; there are other methods; which one's the best depends on a lot of facts that you haven't told us (and might not know). – Gerry Myerson Jul 14 '13 at 03:53
  • 2
    When trying to fit a mathematical model to data (by adjustment of parameters), the choice of mathematical model (your "nonlinear equation") is usually motivated by an understanding of the source of the data (which you have not shared). For example, your second coordinates appear to increase with decreasing first coordinates. Real world applications entail some apriori explanatory framework (theory) which allows us to assess the meaning of such a correlation and then choose a compatible numeric formulation. – hardmath Jul 14 '13 at 04:09

1 Answers1

-1

It's pretty much impossible to come up with "the" nonlinear equation that describes the data without more information on the problem, such as the source of the data and the desired fitted form, or at the very least the units. What's the independent and dependent variables? Time? Position? Elevation? Force? Voltage? Money? You can't just list random data points and expect to automatically fit the "correct" function to it; we need more information to solve this problem.

logosintegralis
  • 330
  • 1
    An Answer should provide more than a complaint that the Question is "pretty much impossible". There are things you can do easily enough to provide value to Readers while scrutinizing the Question's assumptions. E.g. pop the six data points in a spreadsheet and do a linear fit, and show that the need for "nonlinear equation" is doubtful (linear fit is pretty good). – hardmath Jul 14 '13 at 13:20