linear fit in Log Scale + extension

Question

Niklas Kurz 2020-6-16

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/549492-linear-fit-in-log-scale-extension

评论： Niklas Kurz 2020-6-17

Hello, also this question might be trivial and frequently answered, but yet again I'm not able to grasp it. Long story short, that's what I got:

ind =  f>=5000& f <= 10000 
p = polyfit(f(ind),qU(ind),1);
t = polyval(p,(f(ind)));
plot(f(ind),t,'-r','Linewidth',3);

actually this looks like a straight line to me, even if I set:

set(gca,'XScale','log')
set(gca,'YScale','log')

However I want to extend this line, lets say to [5000 15000]. How can I finally achieve this? Thx in advance

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

David Hill 2020-6-16

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/549492-linear-fit-in-log-scale-extension#answer_452223

编辑：David Hill 2020-6-16

在 MATLAB Online 中打开

ind =  f>=5000& f <= 10000 
p = polyfit(f(ind),qU(ind),1);%you will always get a line if degree is 1
x=5000:15000;
t = polyval(p,x);
plot(x,t,'-r','Linewidth',3);

3 个评论
显示 1更早的评论隐藏 1更早的评论

Niklas Kurz 2020-6-16

编辑：Niklas Kurz 2020-6-16

so in comparision: Linear scale with Lines

David Hill 2020-6-17

在 MATLAB Online 中打开

p1 = polyfit(f(1:2),qU(1:2),1);%however many points you want
p2 = polyfit(f(3:end),qU(3:end),1);%second line
x=5000:15000;
t1 = polyval(p1,x);
t2 = polyval(p2,x);
plot(x,t1,'-r',x,t2,'-g','Linewidth',3);

请先登录，再进行评论。

Answer 2

Niklas Kurz 2020-6-17

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/549492-linear-fit-in-log-scale-extension#answer_452595

在 MATLAB Online 中打开

Thank you for your effort, that is what I implemeted:

p1 = polyfit(f(1:2),qU(1:2),1);%however many points you want
x=5000:50000;
t1 = polyval(p1,x);
plot(x,t1,'-r','Linewidth',3);
p2 = polyfit(f(3:8),qU(3:8),1);%second line
x=15000:640000;
t2 = polyval(p2,x);
plot(x,t2,'-g','Linewidth',3);

However: that is what I receive: Sorry for bothering you still, but its not quite running well due to my lack of log-knowledge.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Niklas Kurz 2020-6-17

在 MATLAB Online 中打开

Okay, I finally unravelled he solution, it proved to be a really tricky one:

p1 = polyfit(log(f(1:2)),log(qU(1:2)),1);%however many points you want
x=5000:50000;
t1 = polyval(p1,log(x));
hold on
loglog(x,exp(t1))
p2 = polyfit(log(f(3:8)),log(qU(3:8)),1);%however many points you want
x=5000:640000;
t2 = polyval(p2,log(x));
hold on
loglog(x,exp(t2))