Logarithmic best-fit line

Hi all,

I have two vectors x and y, and I need to add a logarithmic best-fit line to the graph. This is a pretty easy feature to add on excel, but I am having trouble doing it in Matlab. Could someone please write some script using the two vectors on how to accomplish this? (Also, please no links. I tried to follow other Mathworks answers, but I still am not understanding this. Thanks!

Below is what the best fit line looks like on Excel (for reference to what I am asking)

Answer 1

[Walter Roberson]

No. You have negative y data, and logarithmic fit of that would go complex valued.

If you only want the x to be logarithmic, then either use semilogx() or else plot in your normal way and the set the XScale property of the axes to 'log'

[Vladimir Sovkov]

In the graph, the experimental points do not contain negative or zero values in the x-axis, which logariths are real values. Besides, the logarithmic fit is not obligatory the pure logarithm, it can contain intercepts in both x and y, making it possible to work with zero and negative values within the real-type arithmetics.

[Walter Roberson]

When I search around, I do not find the possibility of there being a linear shift in anything called a "logarithmic fit". I only find that the log space might have a linear shift -- which corresponds to a multiplicative constant in linear space.

[Walter Roberson]

https://www.mathworks.com/matlabcentral/fileexchange/29545-power-law-exponential-and-logarithmic-fit