Estimates of parameters =(₁, ₂, ...._n)^T are determined by solving linear least squares problem

min _i=1^m (y_i - c_i^T)²

associated with fitting model to data y=(y_1, y₂, ......y_m), m is equal or larger than n.

Let C be the m x n observation matrix whose i-th row is c^T_i

Assuming the uncertainties associated with the data are such that u (y_i) = , i=1,....,m, write down the expressions for the least square estimate _cx and its associated variance matrix V in the terms of C, y and .