fitting straight line to part of data and extrapolate to x-axis at y = zero

please I want syntax to fit straight line on part of data points which lay on a straight in this graph (I point red circle on this data) , and then extraplotate to X-axis at y=zero in order to calculate X Value at Y=ZERO

Answer 1

[Image Analyst]

If you don't know where the knee/kink/corner of your data is, then you can use a piecewise linear fit to fit the left and right portion to a line. The attached demo will figure out where that knee point is and give you the equation for each line.

[Hazem Mohammed]

Sorry I cannot fit this demo to my model

Can I send you my model and make this model for me

[Image Analyst]

Sure. You can attach your data in a .mat file. What is your criteria for defining where the knee is in that data?

[Hazem Mohammed]

I plot data of X and Y and curve appeared as this so I want to fit straight on the data which I foused on it an then calculate X intersection at y=zero to calculate G_new which must connverge the assumed G with accepted tolerance error (Err) in itterative scheme.

[Image Analyst]

Try this:

[Hazem Mohammed]

thanks so much for your help

but I want some thing in this model, is to iterate on G as G_new is the intersection of X-axis ay y =zero but the problem is G_new = Nan

So please could you help me to iterate on G in oreder to produce G_new with a real value

the second issue is straight line is perfectly as you provide in the model but I want to remove the staraight line which I circled it by green line as attached

Thanks in advance

[Image Analyst]

Add this to the end:

figure;
indexes = x > xc;
% Plot data from the intersection point to the right.
plot(x(indexes), yNoisy(indexes),'b*','LineWidth',2);
% Plot line 2 over/through data.
hold on;
plot(x(indexes), y2Fitted(indexes), 'r-', 'LineWidth', 2);
xlabel('Normalized Cumulative Qn', 'FontSize', fontSize);
ylabel('qg/\Delta p', 'FontSize', fontSize);
grid on;
title('Y vs. X', 'FontSize', fontSize, 'interpreter', 'none');
g = gcf;
g.WindowState = 'maximized'

[Hazem Mohammed]

But the problem still G_new = Nan

[Image Analyst]

That's your code. Note that I did not fit the line the way you did and I did not get an x value of nan when y = 0. So take out your code and replace it with mine.

[Hazem Mohammed]

I'd like to thank you very much , you are solved my problem

Just I adding your code inside iteration loop, so I get G_new at X0 without getting any Nan values.

So, the last issue needing your help : you sent me code for removing the upper straight line, but the problem straight line removed with data points, but I need to only remove upper straight line and keep data points, in other words I want upper points appeared without fitting it to staright line on the same curve so the net result which I need is some thing like that graph "attached"

Many thanks again for your proffisonal helping me

Best Regards

[Image Analyst]

See attached.

[Hazem Mohamed]

Thanks so much

Answer 2

[Matt J]

region=(y<0.01);

p=polyfit( x(region),y(region) ,1);

xroot= roots(p) % the extrapolated root