a) Which of the following algorithms does not suffer from a cold-start problem when a new item is arrived, Content-based or Collaborative filtering based recommender systems? Justify your answer. b) Consider the data given below and compute the information gain if you split the dataset based on the attribute Opponent? Game# Opponent Point Guard Fouls Result 1 Weak Strong No Win 2 Strong Strong Many Loss 3 Strong Weak Many Loss 4 Weak Weak Many Loss 5 Strong Weak No Win 6 Weak Weak Few Win 7 Strong Weak Few Loss 8 Strong Strong Few Win c) Consider the following Bayesian network. Suppose that the grass is wet and it is raining. What is the probability that it is cloudy? F CP(SIC) T .10 .50 Sprinkler Example P(C) 50 Cloudy Wet Grass SRP(WIS,R) T T .99 T F .90 FT .90 F F .01 CP(RIC) Rain T .80 F .20