Weighted Least Square Regression

I have a x matrix with 5 columns and Y matrix with one column of values. I want to minimize this function for all observations:sum(y-sum(alpha*x))^2. Which function can do this thing? I want to have the values of alpha and sum of alpha will be one.

Answer 1

[Star Strider]

I don’t understand your equation. Weighted least squares requires:

WSSCF = sum(w.*(y-f(x)).^2);        % Weighted Least Squares Cost Function

where ‘w’ is the vector of weights, and f(x) is actually a function of ‘x’ that maps ‘x’ to ‘y’. (This is schematic only. The actual function f(b,x) is a function of the parameters ‘b’ as well.

There are several linear and nonlinear parameter estimation functions that can do what you want, but the one you use depends on the nature of your data, the toolboxes you have available, and the model you want to use to fit it.