could you please help me write the solution, thanks!

3. Again, consider a binary classification problem. Now we have N pairs of data points {(x(1), C(1)), ...,(x(N), C(M)) }, where xE Rd is the feature vector and C; is the binary class label (i.e. C; is either 0 or 1). We consider using Logistic Regression for this problem. Suppose y 2) = Wix + wo, (3) the prediction is based on Sigmoid function 1 sigmoid(y ")) = 1 + exp(-y()) (4) If sigmoid(y")) 2 0.5, x is predicted as a data point from class 1; otherwise, class 0. The likelihood function is N L(w1, wo) = p(x()) (1 - p(x(1)))(1-c()) (5) i=1 a. (6 points) Prove that the log-likelihood function can be rewritten as N I(W1, wo) = log(L(w1, wo)) = (wix + wo) - log[1 + exp(wix")+ wo)] (6) i=1 b. (4 points) Compute the derivatives of the log-likelihood function w.r.t. w1 and wo, respectively