Answer the following two questions, thanks!

Given a training set DU = {(23m 6 RD, ya 6 {+1, 1})},\":'=1 for binary classi cation, let 11s t a Naive Bayes model to it; i.e., p(-y|:c) 0c My) Hilptdy}, where p(:c[d]|y) is a Gaussian distribution N(1Ld'y,0'). That is, we assume that for each dimension d, the two onadimensional Gaussian distributions (one for each class) share the same variance 03. For My), let p(+1) = A. Please show 1 . m (y 6 {+1. 1}) [10 po1nts]; 1. b ' 'tt p(y|:c] can e re \"r1 en as 1+ eM 2. what the corresponding to and b are [5 points]. Your answers should be based on the notations his}, 5rd, and A. For to E R9 , you may simply write the expression of HIM for a specic (1. Similar to the HW #2 GDA question, this question shows you that under certain assumptions, Naive Bayes will lead to a linear classier. Please write down your derivation with no more than 15 lines for each answer. Hint1: You do not need to perform maximum likelihood estimation. You can assume that p.13, and 5rd are given and xed. Hint2: Mm) = My = +1,:c) + My = 1,;c), which comes from the marginal probability