linear fit of a semilog graph

Hello,

I have my data as follows with F1, F2, F3, N1, N2 and N3. I want to do a linear fit of my data and plot that. I tried polyfit as seen in my code. But it just plots a horizontal line...? I tried adding my vectors to a Ntot and Ftot as seen.

What should I do? Thank you.

F1=[200 200 200];

N1=[10000 15000 20000];

F2=[300 300 300];

N2=[8000 7000 7500];

F3=[100 100 100];

N3=[120000 140000 80000];

Ftot=[F3 F2 F1];

Ntot=[N3 N2 N1];

figure

semilogx(N1,F1,'o')

hold on

semilogx(N2,F2,'o')

semilogx(N3,F3,'o')

xlabel('N');

ylabel('F');

grid on

P = polyfit(Ntot,Ftot,1);

    yfit = P(1)*Ftot+P(2);

    plot(Ntot,yfit,'r-.');

xlim([1000 500000])

ylim([0 max(Ftot)+50])

[Torsten]

According to your call to polyfit, the next line must read

yfit = P(1)*Ntot+P(2);

instead of

yfit = P(1)*Ftot+P(2);

Best wishes

Torsten.

[Star Strider]

The other option, of course, is to use the polyval function.

Answer 1

[Erik]

You should sort the values in

Ntot

using

[Ntot,i] = sort(Ntot);

and sort

Ftot

accordingly using

Ftot = Ftot(i);

Then you can do the

polyfit(...)

and as Torsten mentions, you should use

yfit = P(1)*Ntot+P(2);

Then plot it using

semilogx(Ntot,yfit,'r-')

although

plot(...)

also works. If you want the semilogarithmic plot to have a straight line, then use

log10(Ntot)

instead of

Ntot

in the line containing the

polyfit(...)

and the line containing

yfit = ...

[Daniel]

Thank you. Okey, but now I have a plot like the attached picture.

Can I do so a straight line goes from the data at 300 at the y-axis to the data at 100 on the y-axis?

Is it some way to do an approximation for that?