A general single-equation nonlinear model may be expressed in matrix notation as 
where Yc is the model vector containing the computed y; (calc), [F(xj)]' is the transpose of a column matrix expressing the functionality of the model, and b is the vector containing the values of the model parameters. If Y is the matrix containing the values of yy(meas), the unweighted error sum S becomes
A detailed discussion of parameter estimation by matrix methods can be found in the book by Bates and Watts . Bevington's book  provides bridges between the algebraic and matrix approaches.
Was this article helpful?