Problem 9. (16 pts) A training set for a classification task is provided below. The dataset has two features and one categorical label, y, which has two possible classes: "red" and "blue". You are asked to score a logistic regression model for this training data Round all answers to three decimal places on this problem. a) x(1) 1 5 6 2 7 x(2) 6 3 5 2 1 y red blue blue red red In the model below, is an estimate for the probability that an observation falls into the red class. Let f = p for red observations and let it = 1 - for blue observations. For the following model, find and it for each observation. 1 = 1+-4+0.6x(1)+0.5x(2) b) Calculate the negative log-likelihood score for this model. c) When calculating likelihood scores, why do we typically work with log-likelihood rather than likelihood?