To take parameter of a fit from one section to use in another fit in the next section.
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
aneps
am 21 Mai 2014
Kommentiert: Sean de Wolski
am 22 Mai 2014
I have a program which make a linear fit on a data. I want to use the slope of this linear fit on another data where I fit exponential (a*exp(b*x)) as b. Part of my program which do this is as below:
%%SECTION 1:THIS PART FIT LINEAR FIT ON THE PLOT.
[FTLx,FTLy]=ginput(2);
indexL=find(B(:,1)>=FTLx(1) & B(:,1)<=FTLx(2));
NewXL=B(B(:,1)>=FTLx(1) & B(:,1)<=FTLx(2));
NewYL=B(indexL,2);
P=polyfit(NewXL,log(NewYL),1);
LFit=P(1)*NewXL+P(2);
slope=P(1) %THIS slope I USE IN THE NEXT SECTION
hold on
figure(3)
subplot(1,2,1),plot(NewXL,LFit,'-b',NewX,log(NewY),':or')
%%SECTION 2: THIS SECTION FITS EXPONENTIAL... THIS SECTION USE 'slope' FROM THE PREVIOUS SECTION
options = fitoptions;
options.Normal = 'on';
ft=fittype('a*exp(slope*x)','independent','x','dependent','y');%THIS 'slope' SEEMS NOT TAKING THE VALUE FROM THE PREVIOUS SECTION!! :( :( :(
[fitresult, gof]=fit(NewX,NewY,ft,options)
figure(3)
subplot(1,2,2), plot(fitresult,'-b',NewX,NewY,':or')
title('Exponential Fit in the selected range')
Here in the second section, I am using 'slope' from the first section. But it seems to taking the value of 'slope' from the first section! Is there anything wrong? How can I correct it?
1 Kommentar
dpb
am 21 Mai 2014
I've never actually used the wrapper stuff but looks like perhaps if you were to wrap in an anonymous function --
ft = fittype( @(a,x) a*exp(slope*x), 'problem', 'slope' );
f1 = fit(NewX, NewY, ft, 'problem', slope);
might do the trick...
Akzeptierte Antwort
Sean de Wolski
am 21 Mai 2014
% Sample data
x = (1:10).';
y = cumsum(rand(10,1));
% First fit
ft = fittype(@(m,b,x)m.*x+b);
fr = fit(x,y,ft)
% Second fit
ft2 = fittype(@(a,x)a*exp(fr.m*x));
fr2 = fit(x,y,ft2)
Using anonymous functions. You could do this with strings too but it would involve some kludge with sprintf / num2str etc.
2 Kommentare
Sean de Wolski
am 22 Mai 2014
You need to pass x into the fit result object:
plot(x,y,'b*',x,fr(x),'r-')
Weitere Antworten (0)
Siehe auch
Kategorien
Mehr zu Linear and Nonlinear Regression finden Sie in Help Center und File Exchange
Produkte
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!