Could anyone help me with this? I will definitely thumb up.

1. Prove the update guarantee of AdaBoost, that is: show that the base learner ht\" selected at round t + 1 cannot he t under the weak learner assumption, where ht he the base learner selected at round t. 2. Consider a case where we have learned a conditional probability distribution P[ylx). Suppose there are onl}F two classesI and let p0 = Ply = U'Ix} and p1 = P[y = 1|x}. Consider the loss matrix below: eeeeeeeee leee e n \"Ii-n -n Show that the decision 3,7 that minimizes the expected loss is equivalent to setting a probability threshold 6' and predicting g} = CI if 331 a: {'3 and g} = 1 if 131 1'3 {'3 . What is {'3 as a function of {11 and a2? 3. lGiven two probability mass functions {pi} and {as}, then 2 a 111 E 3 it i 11-: with equality if and onl}F if q; = 11;, for all