I have doubts in the following questions:
Q3) Shouldn’t the weights be n for n features? Are we considering bias as the weight for dummy feature x0, in this question?
Q4) We maximize the log likelihood to find the best value of theta so shouldn’t large values of the log-likelihood statistic indicate that the statistical model fits the data well?
Q6) The diagrams in options (b) and © are same.
Hey @SanchitSayala,
Q3) Yes, n weights are taken for n features and +1 for the bias term. So a total of (features + 1) weights will be learned by our model.
Q4) No it follows that large values of the log-likelihood statistic indicate poorly fitting statistical models, because the larger the value of the log-likelihood, the more unexplained observations there are.
Q6) Yes you can leave that question. Although the correct answer to that question is option B only.
Hope this helps.
Happy Learning 