Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Fitting an inverse quadratic curve

  1. Apr 10, 2006 #1
    If I know 2 points in the +x,+y quadrant, and I want to find the

    Code (Text):

    f(x) =    a
    curve that passes through both points (a and b are constants).

    This is probably either really simple or impossible. :rolleyes:
  2. jcsd
  3. Apr 10, 2006 #2
    Yes... this is either really simple or impossible.

    Let [tex]k=\frac{a}{b^2}[/tex]. Since [tex]f(x) = \frac{a}{(bx)^2} = \frac{a}{b^2x^2}=\frac{\frac{a}{b^2}}{x^2}[/tex], we can say [tex]f(x) = \frac{k}{x^2}[/tex].

    If the two points are given by [tex]P_1 = (x_1,y_1)[/tex] and [tex]P_2 = (x_2,y_2)[/tex], then they both lie on such a curve if and only if [tex]y_1 = \frac{k}{x_1^2}[/tex] and [tex]y_2 = \frac{k}{x_2^2}[/tex]. That is, if [tex]x_1^2y_1=x_2^2y_2[/tex]. If this is the case, then let [tex]k=x_1^2y_1[/tex], and translate to a and b. Since we've got two variables and one equation, we're free to choose one of them. For ease, let's make b=1. Then a=k. So, [tex]f(x)=\frac{k}{x^2}[/tex].
  4. Apr 11, 2006 #3
    Though I can't think how I missed that but I do feel quite stupid now :yuck:
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Fitting an inverse quadratic curve
  1. Fitting A Curve (Replies: 1)

  2. Curve fitting (Replies: 0)