Hi, I'd like to adjust a set of points to an expression like this "E=c1*H+c2​*H*T^(-2)+​c3*T", please help!!...Thanks

"E=c1*H+c2*H*T^(-2)+c3*T"

where

E,H,T are variables c1, c2, c3 are the regression coefficients

I have all the data necessary of the three variables from a 1 month period, I'd like to know how to obtain the regression coefficients I think that it's a multiple non linear regression but, I don't know what functions of matlab I could use, I'll be thankful for your help.

Thank you very much!!

Answer 1

[Brian B]

This is a linear regression problem, since it is linear in the coefficients. Assuming E, H, and T are column vectors, you might simply try:

   R = [H H.*T.^(-2) T];
   c = R\E;

This finds the least-squares solution for the linear equation E = R*c.

Answer 2

[Star Strider]

See if this example does what you want:

H = randi(10, 5, 1);        % Define H
T = randi(10, 5, 1)+10;     % Define T
c = [3 5 7]';               % Define c

X = [H  H./(T.^2)  T];      % Design matrix
Y = X*c + 0.1*rand(5,1);    % Create dependent variable + noise

c = X\Y                     % Solve for coefficients

If you have the Statistics Toolbox, see the documentation for the LinearModel class.