Is there a way to add linear fit equations and r^2 values to multiple plots?

I have six subplots. Each is a property-property plot of pH and temperature. I would like to add the linear fit equation and r^2 value to each plot. I have used the Basic Fitting toolbox in the past to display linear fits on one plot at a time and to show the equation. I am wondering if I can add a line (or lines) to my plotting code to do this. Please see my plot (currently with one linear fit equation shown). '

Thank you so much!

Answer 1

[Star Strider]

There are many ways to do this.

Here is one:

    function [slope,intcpt,Rsq] = linfit(x,y)
        B = [x(:) ones(size(x(:)))] \ y(:);
        ypred = [x(:) ones(size(x(:)))] * B
        SSE=sum((y(:)-ypred).^2)
        SST=sum((y(:)-mean(y)).^2)
        Rsq   = 1 - (SSE/SST);
        slope = B(1);
        intcpt = B(2);
    end

X = rand(1, 20);                                                % ‘X’
Y = 2*X+1 + randn(1,20)*0.1;                                    % ‘Y’

[slope,intcpt,Rsq] = linfit(X,Y)

plot(X,Y,'.')
hold on
Ln = [min(X) 1; max(X) 1] * [slope; intcpt];
plot([min(X) max(X)], Ln', '-r')
hold off
text(0.7, 0.8, sprintf('R^2 = %.3f', Rsq))

The code outside the function tests it and demonstrates its use. If you want to use it, save it as: linfit.m on your MATLAB user path, then call it as I did in the demonstration code.

[Heidi Hirsh]

Thank you! This works for finding the linear fit equation and r^2. But when I run your code I don't think the text is plotting on the figure. See my figure attached.

[Heidi Hirsh]

This is how I'm coding my own figure if it helps:

f1=figure(1)

subplot(3,2,1)

plot(PK.PH(2,:),PK.TEMP(2,:),'.','color','k')

legend('6.1 mab')

ylim([10 18])

xlim([7.5 8.4])

set(gca,'fontsize',14)

ylabel('Temperature')

xlabel('pH')

subplot(3,2,2)

plot(PK.PH(3,:),PK.TEMP(3,:),'.','color','k')

legend('5.2 mab')

ylim([10 18])

xlim([7.5 8.4])

set(gca,'fontsize',14)

ylabel('Temperature')

xlabel('pH')

subplot(3,2,3)

plot(PK.PH(4,:),PK.TEMP(4,:),'.','color','k')

legend('4.4 mab')

ylim([10 18])

xlim([7.5 8.4])

set(gca,'fontsize',14)

ylabel('Temperature')

xlabel('pH')

subplot(3,2,4)

plot(PK.PH(5,:),PK.TEMP(5,:),'.','color','k')

legend('3.3 mab')

ylim([10 18])

xlim([7.5 8.4])

set(gca,'fontsize',14)

ylabel('Temperature')

xlabel('pH')

subplot(3,2,5)

plot(PK.PH(6,:),PK.TEMP(6,:),'.','color','k')

legend('2.6 mab')

ylim([10 18])

xlim([7.5 8.4])

set(gca,'fontsize',14)

ylabel('Temperature')

xlabel('pH')

subplot(3,2,6)

plot(PK.PH(7,:),PK.TEMP(7,:),'.','color','k')

legend('1.5 mab')

ylim([10 18])

xlim([7.5 8.4])

set(gca,'fontsize',14)

ylabel('Temperature')

xlabel('pH')

[Star Strider]

My pleasure!

The code I wrote was for my demonstration data. A more robust text call is:

text(min(xlim)+0.7*diff(xlim), min(ylim)+0.2*diff(ylim), sprintf('R^2 = %.3f', Rsq))

That should work for all your plots. Adjust the constants (0.7 and 0.2 here) to place it where you want it. It should be in about the same relative position in every plot.

I also just discovered that I omitted some semicolons in ‘linfit’. It should be:

    function [slope,intcpt,Rsq] = linfit(x,y)
        B = [x(:) ones(size(x(:)))] \ y(:);
        ypred = [x(:) ones(size(x(:)))] * B;
        SSE=sum((y(:)-ypred).^2);
        SST=sum((y(:)-mean(y)).^2);
        Rsq   = 1 - (SSE/SST);
        slope = B(1);
        intcpt = B(2);
    end

I was verifying it to be certain it did what I wanted it to do, and forgot to put them back before I posted it.