Modify the RSS function

[justinlin]

The current RSS function in datastruct contains the following prediction of y and p, which is wrong: y = -pinv(A)P(:,i); p = -AY(:,i); We should reimplement RSS calculation such that the method can take in the predicted Y and P, can do balancing of the errors between Y and P, calculate both weighted and unweighted RSS.

[Andreas Tjärnberg]

Please explain why the current way of predicting y and p are wrong. It might be worth making an RSS methods that can balance errors.