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

Model: y = a + b*Ln(x)+ c*x

  1. Jan 18, 2010 #1
    I have calibration data and I'm want to use the following model:

    y = a + b*Ln(x)+ c*x

    I'm looking for a way to make initial guesses for the three paramaters a, b and c.

    Is there any way to linearize this expression so that I can use least squares analysis to determine values for a, b and c.
     
  2. jcsd
  3. Jan 18, 2010 #2
    Re: Linearize

    Your problem is already linear in a, b and c (which are the variables you're trying to fit).

    edit: to be a bit more explicit, given a pair of data values (x_i, y_i) you get a the following equation

    a*1 + b*Ln(x_i) + c*x_i = y_i

    which is a linear combination of the unknowns (a, b, c).
     
    Last edited: Jan 18, 2010
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook