Extend a line made with polyfit and polyval

18 次查看(过去 30 天)
Hi
I have a problem with extending my polyfit line. I tried to extend it by changing "T" for polyval (which is temperature and x-axis) to contain the whole dataset for temperature to get an extended line ranging the whole plot. I wanted a line that was extended like 4 degrees celsius, but wanted to checking if it worked by including the whole data set first. It did not. All the y-values decreased and the extended line does not look linear any more.
Should this solution work, and if so what have I done wrong?
Is there another solution to extend the line?
Thanks a lot for the help
% Getting index for temperature values
[d,posmin] = min(abs(x11-0));
[d,negmax] = min(x11-10);
x111 = x11(negmax:lastrow11);
[d,negmin1] = min(abs(x111-0));
negmin = negmax+negmin1-1;
% Getting x and y values
T11 = x11(1:posmin); A11 = y11(1:posmin);
T12 = x11(posmin:negmax); A12 = y11(posmin:negmax);
T13 = x11(negmax:negmin); A13 = y11(negmax:negmin);
T14 = x11(negmin:lastrow11); A14 = y11(negmin:lastrow11);
% Fitting values
[p11,s11] = polyfit(T11,A11,1);
f11 = polyval(p11,wholedataset);
[p12,s12] = polyfit(T12,A12,1);
f12 = polyval(p12,T12);
[p13,s13] = polyfit(T13,A13,1);
f13 = polyval(p13,T13);
[p14,s14] = polyfit(T14,A14,1);
f14 = polyval(p14,T14);

采纳的回答

Star Strider
Star Strider 2018-12-17
You are doing linear regressions, so extending the line is not a problem. (It easily could be a problem with higher-order polynomials.)
It is not necessary to do new regressions to extend the line. Just the same polynomial values with new limits:
p11 = polyfit((1:10)', 2*(1:10)'+randn(1,10)', 1); % Original Linear Regressopm
orig = polyval(p11, (1:10)'); % Evaluated
xtnd1 = polyval(p11, [-5 15]); % Extend Using ‘p11’ With New Limits
figure
plot((1:10), orig, 'b-', [-5 15], xtnd1, '--r')
grid

更多回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Line Plots 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by