Please help.

Consider a Bayesian classifier, with distributions being n-dimensional multivariate Bernoulli random variables. The conditional probability for a given category is n P(x|0) = 107 (1 -0;)1-2, i=1 and let D = {x1, . .. xx} be a set of k samples independently drawn from this proability density. (a) If s = ($1, . .., Sn) is the sum of the k samples, show that n P(DIO) = I1 0(1 - 0;) k-si i=1 (b) Assuming a uniform prior distribution for 0 show that n P(0 D) = II (k + 1)! i=1 Si!(k - s;)!vil ( 1 - 0;) k-si, (c) Plot this density for the case n = 1, k = 1 and for the two resulting possibilities for si. (d) Integrate the product P(x|0) P(0|D) over 0 to obtain n P(x|D) = II Si + 1 \\ Ci Si + 1 \\ 1-xi i=1 k + 2 k + 2 (e) Compare the expressions for P(x|D) and P(x|0). What is 0, the effective Bayesian estimate for 0