Extract data using ginput and fit with a formula of your choice

I am choosing two end points using ginput [x,y] = ginput(2) and then I need to separate my data set which lies between the two end points. The separated data set must be fitted into the following formula.

y(1)*exp(-(x-x(1))/T)+y(2)

y(1),x(1),y(2) are the values taken from ginput. T is a parameter that I need to determine.

How can be done?

[Mathieu NOE]

hello

have you started coding something ?

[C A]

Yes, I already have a plot. I just need to fit the curve between the two points that I choose, instead of fitting the entire data set.

[Mathieu NOE]

hello

maybe if you share the code and data I can work on it

Answer 1

[JJ Lamb]

I'm going to say that the input data for the graph is in a 2 column vector called "myData"

[x,y] = ginput(2);

% assuming you need points based on where you clicked on the x axis
xData = myData(x(1):x(2),1);
yData = myData(x(1):x(2),2); 

myFun = @(T,x) y(1)*exp(-(x-x(1))/T)+y(2); % copied your function

TT = lsqcurvefit(myFun,1,xData,yData); % TT is the result of the curvefit
% the second input 1 is just an initial guess.
% might need to adjust it to a more reasonable starting point for your data

% Plot your results
figure
plot(xData,yData,'rx'); % original data
hold on
xx = min(xData):max(xData)-min(xData)/100:max(xData); % dummy variable for plotting
plot(xx,myFun(TT,xx),'b')

[C A]

Thank you, may I ask why you have used 2 here (yData = myData(x(1):x(2),2)) instead of 1 ?

[JJ Lamb]

I assumed that the y data was in the second column. Column 1 is x data, column 2 is y data. You may have them in separate variables already, in that case you would just need:

xData = oldX(x(1):x(2));
yData = oldY(x(1):x(2));