That's a rough measure of the a priori uncertainty. But you want the uncertainty conditioned to the particular observations you have made. A ton more here: h

That's a rough measure of the a priori uncertainty. But you want the uncertainty conditioned to the particular observations you have made. A ton more here: http://mlg.eng.cam.ac.uk/yarin/blog_2248.html http://mlg.eng.cam.ac.uk/yarin/blog_2248.html