How to make linear fit?
Show older comments
Hi; I want to plot a linear fit for the graphic in the attached file. And also I want to calculate slope of this linear fit? How can I do this? File is in the attached file.. Thanks a lot.
Answers (2)
Star Strider
on 22 Jun 2018
Try this:
I = openfig('test.fig');
Ax = gca;
L1 = findobj(Ax, 'Type','line'); % Find ‘line’ Object
x = L1.XData; % Extract X-Data
y = L1.YData; % Extract Y-Data
XL = [min(x) max(x)]; % Get X-Limits
P = [x(:), ones(size(x(:)))] \ y(:); % Estimate Linear Fit Parameters
LinFit = [XL(:) [1;1]] * P; % Linear Fit
Slope = P(1);
hold on
plot(XL, LinFit, '-r', 'LineWidth',1.5)
hold off
5 Comments
Steven Lord
on 22 Jun 2018
Matt Dickson
on 22 Jun 2018
Also consider using polyfit to perform the regression and simplify the code, if you want it done in the script.
Star Strider
on 22 Jun 2018
My code estimates, calculates, and plots a linear fit.
It estimates the parameters as ‘P’, and then calculates a regression line in ‘LinFit’. This is likely more efficient than polyfit and polyval for a simple linear fit.
nancy
on 22 Jun 2018
Star Strider
on 22 Jun 2018
Edited: Star Strider
on 24 Jun 2018
To use polyfit:
I = openfig('test.fig');
L = findobj(gca, 'Type','line'); % Find ‘line’ Object
x = L(2).XData; % Extract X-Data
y = L(2).YData; % Extract Y-Data
x = x(y > -2); % Remove Outliers
y = y(y > -2); % Remove Outliers
XL = [min(x) max(x)]; % Get X-Limits
P = polyfit(x, y, 1); % Estimate Linear Fit Parameters
LinFit = polyval(P, XL); % Linear Fit
Slope = P(1);
figure
plot(x, y)
hold on
plot(XL, LinFit, '-r', 'LineWidth',1.5)
hold off

EDIT — Added plot image.
Image Analyst
on 23 Jun 2018
0 votes
See my attached polyfit() demo.

2 Comments
nancy
on 24 Jun 2018
Image Analyst
on 24 Jun 2018
The demo runs fine. How did you change it? Post your data and code.
Categories
Find more on Linear and Nonlinear Regression in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!