In weighted least-squares linear regression, we have weight ri corresponding to each data measurement. Our goal is to fit data points in proportion to their weights by minimizing following objective function:
E(w) = sum from i=1 to m [r(i) (y(i) - (wx(i) + b))2]
where w and b are model parameters, training data are pairs {x(i), y(i)}, i = 1, ..., m.
Deduce closed-form expression for estimates of w and b which minimize objective function.