Thanks. That post took at least 10 hours over the past few weeks to write; it's gratifying to know you found it useful. The next one will probably be on regula

Thanks. That post took at least 10 hours over the past few weeks to write; it's gratifying to know you found it useful. The next one will probably be on regularized least squares (i.e. choosing an x that is close to optimal and also small, so it behaves well in later calculations). You might also look at Stephen Boyd's videos on linear estimation: http://www.youtube.com/view_play_list?p=06960BA52D0DB32B http://www.youtube.com/view_play_list?p=06960BA52D0DB32B I think least squares is covered in lectures 5 and 6.