extrapolate data to cross x axis

I have this curve and need to extrapolate it so both ends cross x axis, I have used Vq = INTERP1(X,V,Xq,'linear",'extrap') and tried other 'method's but it did not look "good".

any suggestions how to do it nicely? thanks.

Answer 1

[Star Strider]

There are 19 variables in that .mat file. I have no idea which ones are to be extrapolated, so I took a wild guess on two of them.

Try these —

LD = load('QTHLS.mat');

figure
plot(LD.MHLSx, LD.MHLS)
grid
MHLS1i = interp1(LD.MHLS(1:10), LD.MHLSx(1:10), 0, 'linear','extrap')
p = polyfit(LD.MHLSx(end-9:end), LD.MHLS(end-9:end),1)
MHLS2i = -p(2)/p(1)
hold on
plot([MHLS1i, LD.MHLSx(1)], [0 LD.MHLS(1)], '-r')
plot([MHLS2i, LD.MHLSx(end)], [0 LD.MHLS(end)], '-r')
hold off

figure
plot(LD.XLSx, LD.XLS)
grid
XLS1i = interp1(LD.XLS(1:10), LD.XLSx(1:10), 0, 'linear','extrap')
p = polyfit(LD.XLSx(end-9:end), LD.XLS(end-9:end),1)
XLS2i = -p(2)/p(1)
hold on
plot([XLS1i, LD.XLSx(1)], [0 LD.XLS(1)], '-r')
plot([XLS2i, LD.XLSx(end)], [0 LD.XLS(end)], '-r')
hold off

The online Run feature does not work well with .mat files, so I did this offline.

They also demonstratae two diferent methods of doing the extrapolation.

.

Answer 2

[Max Heimann]

In order for this question to be answered you should define what you mean by "nice". Unfortunately your code sample is not runable with the given workspace. But if it were it would still be unclear what would need to change for you to have an acceptable curve.

You could try polyfit and polyval to generate a polynomial which matches your data and then extrapolate by evaluating it outside of your datapoints. This works best if you have some idea how your curve should look like outside of your existing datapoints.

Answer 3

[Asliddin Komilov]

sorry for the mess, I thought I have saved one variable. This is exactly what I needed.

thank you so much

[Star Strider]

As always, my pleasure!