Problem with Power Regression

on 19 Jul 2019
Latest activity Commented on by the cyclist

on 20 Jul 2019

the cyclist (view profile)

I have a coding to develop a formula using power regreesion
Here my coding:
plot(Xbin,Ybin,'+r'), hold on
p = polyfit(log(Xbin),log(Ybin),2);
m = p(1);
b = exp(p(2));
ezplot(@(x) b*Xbin.^m,[Xbin(1) Xbin(end)])
Error in Power_Regression (line 242)
ezplot(@(x) b*Xbin.^m,[Xbin(1) Xbin(end)])
I appreciate the help

Image Analyst

Image Analyst (view profile)

on 19 Jul 2019
Why not use fitnlm() for a better fit? I'm attaching several examples. Would you consider that (fitnlm) instead of taking the log and doing polyfit() on the logged data?

on 20 Jul 2019
Dear Cyclist,

the cyclist (view profile)

on 19 Jul 2019
Edited by the cyclist

the cyclist (view profile)

on 19 Jul 2019

polyfit(...,2) is fitting a constant, linear, and quadratic term. Notice that the output vector p has three elements. You want this instead:
Xbin = (1:10)';
Ybin = 2 * Xbin.^3 + 0.07*randn(10,1);
figure
plot(Xbin,Ybin,'+r'), hold on
p = polyfit(log(Xbin),log(Ybin),1);
m = p(1);
b = exp(p(2));
ezplot(@(x) b*x.^m,[Xbin(1) Xbin(end)])
I also corrected the fact that you used @(Xbin) instead of @(x) in the function definition.
[Sorry if you saw some of the intermediate edits I did, before I realized the true source of your problem.]

on 20 Jul 2019
Thank you, my coding is working know. I want to know how to write the euation and R square inside the figure?
Thank you