how to fit y=ax^b without curve fitting

how to fit a set of data using 𝑦 = 𝑎𝑥^b (logarithmic transformation) ?

[David Goodmanson]

Hi Moza,

you can certainly take the log of both sides and try fitting the straight line

log(y) = c + b*log(x)     % c = log(a)

if that's what you mean.

[Moza Salem]

Yes!!! Thank you so much

Answer 1

[Image Analyst]

You can use fitnlm() - Fit Non-linear model. See attached for a bunch of demos where I use it on a wide variety of equations. Adapt as needed.