how to make linear fit

Hi;

How to make a plot similar to the attached figure ??

Answer 1

[William Rose]

@ahmad Saad,

There are a number of ways you could do it. here is one example:

x=0:100;
y=2*x+20+10*randn(1,101);
p=polyfit(x,y,1);
% Compute fitted values
yfit=polyval(p,x);
% Plot results
plot(x,y,'b.',x,yfit,'-r')

Try it. Good luck.

[William Rose]

@ahmad Saad,

If you also want the RMSE and Pearson correlation, do the same code as before plus a few lines:

x=0:100;
y=2*x+20+10*randn(1,101);
p=polyfit(x,y,1);
rho=corr(x',y')

rho = 0.9876

% Compute fitted values
yfit=polyval(p,x);
rmse=sqrt(mean((yfit-y).^2))

rmse = 9.0371

% Plot results
plot(x,y,'b.',x,yfit,'-r')

And if you want to add the rho and rmse to the plot, do this:

hold on;
s1=sprintf('RMSE=%.2f',rmse);
text(10,200,s1);
s2=sprintf('rho=%.3f',rho);
text(60,200,s2)
text();

OK